U.S. patent application number 12/780296 was filed with the patent office on 2011-05-12 for methods for detection of genetic disorders.
This patent application is currently assigned to RAVGEN, INC.. Invention is credited to Ravinder S. Dhallan.
Application Number | 20110111971 12/780296 |
Document ID | / |
Family ID | 35375636 |
Filed Date | 2011-05-12 |
United States Patent
Application |
20110111971 |
Kind Code |
A1 |
Dhallan; Ravinder S. |
May 12, 2011 |
METHODS FOR DETECTION OF GENETIC DISORDERS
Abstract
The invention provides a method useful for detection of genetic
disorders. The method comprises determining the sequence of alleles
of a locus of interest, and quantitating a ratio for the alleles at
the locus of interest, wherein the ratio indicates the presence or
absence of a chromosomal abnormality. The present invention also
provides a non-invasive method for the detection of chromosomal
abnormalities in a fetus. The invention is especially useful as a
non-invasive method for determining the sequence of fetal DNA. The
invention further provides methods of isolation of free DNA from a
sample.
Inventors: |
Dhallan; Ravinder S.;
(Bethesda, MD) |
Assignee: |
RAVGEN, INC.
Columbia
MD
|
Family ID: |
35375636 |
Appl. No.: |
12/780296 |
Filed: |
May 14, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
11648778 |
Dec 28, 2006 |
7718370 |
|
|
12780296 |
|
|
|
|
10661165 |
Sep 11, 2003 |
7332277 |
|
|
11648778 |
|
|
|
|
PCT/US03/06198 |
Feb 28, 2003 |
|
|
|
10661165 |
|
|
|
|
10093618 |
Mar 11, 2002 |
6977162 |
|
|
PCT/US03/06198 |
|
|
|
|
PCT/US03/27308 |
Aug 29, 2003 |
|
|
|
10661165 |
|
|
|
|
10376770 |
Feb 28, 2003 |
7208274 |
|
|
PCT/US03/27308 |
|
|
|
|
10093618 |
Mar 11, 2002 |
6977162 |
|
|
10376770 |
|
|
|
|
60378354 |
May 8, 2002 |
|
|
|
60360232 |
Mar 1, 2002 |
|
|
|
60378354 |
May 8, 2002 |
|
|
|
60360232 |
Mar 1, 2002 |
|
|
|
Current U.S.
Class: |
506/9 |
Current CPC
Class: |
C12Q 1/683 20130101;
C12Q 1/683 20130101; C12Q 1/6806 20130101; C12Q 1/683 20130101;
C12Q 2525/131 20130101; C12Q 2545/114 20130101; C12Q 2525/131
20130101; C12Q 2535/125 20130101; C12Q 2525/131 20130101; C12Q
2533/101 20130101; C12Q 2521/313 20130101; C12Q 2521/313 20130101;
C12Q 2531/113 20130101; C12Q 2533/101 20130101; C12Q 2527/125
20130101; C12Q 2600/156 20130101; C12Q 2521/313 20130101; C12Q
1/6858 20130101; C12Q 1/6869 20130101; C12Q 1/6858 20130101; C12Q
1/6806 20130101; C12Q 1/683 20130101; C12Q 1/6869 20130101 |
Class at
Publication: |
506/9 ;
435/6 |
International
Class: |
C40B 30/04 20060101
C40B030/04; C12Q 1/68 20060101 C12Q001/68 |
Claims
1. A method for determining a sequence of a locus of interest
comprising: (a) amplifying a locus of interest on template DNA,
wherein the template DNA comprises a mixture of fetal DNA and
maternal DNA and is obtained from a sample from a pregnant female,
using a first and second primers, wherein the second primer
contains a recognition site for a restriction enzyme such that
digestion with the restriction enzyme generates a 5' overhang
containing the locus of interest; (b) digesting the amplified DNA
with the restriction enzyme that recognizes the recognition site on
the second primer; (c) incorporating a nucleotide into the digested
DNA of (b) by using the 5' overhang containing the locus of
interest as a template; and (d) determining the sequence of the
locus of interest by determining the sequence of the DNA of
(c).
2. The method of claim 1, wherein said sample is selected from the
group consisting of blood, serum, plasma, saliva, urine, tears,
vaginal secretion, sweat, lymph fluid, cerebrospinal fluid, mucosa
secretion, peritoneal fluid, ascitic fluid, fecal matter, and body
exudates.
3. The method of claim 1, wherein said sample is blood.
4. The method of claim 3, wherein said blood is obtained from a
human pregnant female when the fetus is at a gestational age
selected from the group consisting of 0-4, 4-8, 8-12, 12-16, 16-20,
20-24, 24-28, 28-32, 32-36, 36-40, 40-44, 44-48, 48-52, and more
than 52 weeks.
5. The method of claim 3, wherein said template DNA is obtained
from plasma from said blood.
6. The method of claim 3, wherein said template DNA is obtained
from serum from said blood.
7. The method of claim 1, wherein the restriction enzyme cuts DNA
at a distance from the recognition site.
8. The method of claim 1, wherein the recognition site is for a
Type IIS restriction enzyme.
9. The method of claim 8, wherein the Type IIS restriction enzyme
is selected from the group consisting of: Alw I, Alw26 I, Bbs I,
Bbv I, BceA I, Bmr I, Bsa I, Bst71 I, BsmA I, BsmB I, BsmF I, BspM
I, Ear I, Fau I, Fok I, Hga I, Pie I, Sap I, SSfaN I, and Sthi32
I.
10. The method of claim 1, wherein said method of amplification is
PCR.
11. The method of claim 10, wherein an annealing temperature for
cycle 1 of PCR is about the melting temperature of the portion of
the 3' region of the second primer that anneals to the template
DNA.
12. The method of claim 11, wherein an annealing temperature for
cycle 2 of PCR is about the melting temperature of the portion of
the 3' region of the first primer that anneals to the template
DNA.
13. The method of claim 12, wherein an annealing temperature for
the remaining cycles of PCR is at about the melting temperature of
the entire second primer.
14. The method of claim 1, wherein determining the sequence
comprises a method selected from the group consisting of allele
specific PCR, mass spectrometry, hybridization, primer extension,
fluorescence resonance energy transfer (FRET), sequencing, Sanger
dideoxy sequencing, DNA microarray, southern blot, slot blot, dot
blot, and MALDI-TOF mass spectrometry.
15. A method for determining a sequence of alleles of a locus of
interest comprising: (a) amplifying alleles of a locus of interest
on a template DNA, wherein the template DNA comprises a mixture of
fetal DNA and maternal DNA and is obtained from a sample from a
pregnant female: using a first and second primers, wherein the
second primer contains a recognition site for a restriction enzyme
such that digestion with the restriction enzyme generates a
5'overhang containing the locus of interest; (b) digesting the
amplified DNA with the restriction enzyme that recognizes the
recognition site on the second primer; (c) incorporating
nucleotides into the digested DNA of (b), wherein; (i) a nucleotide
that terminates elongation, and is complementary to the locus of
interest of an allele, is incorporated into the 5' overhang of said
allele, and (ii) a nucleotide complementary to the locus of
interest of a different allele is incorporated into the 5' overhang
of said different allele, and said terminating nucleotide, which is
complementary to a nucleotide in the 5' overhang of said different
allele, is incorporated into the 5' overhang of said different
allele; (d) determining the sequence of the alleles of a locus of
interest by determining the sequence of the DNA of (c).
16. The method of claim 15, wherein said sample is selected from
the group consisting of blood, serum, plasma, saliva, urine, tears,
vaginal secretion, sweat, lymph fluid, cerebrospinal fluid, mucosa
secretion, peritoneal fluid, ascitic fluid, fecal matter, and body
exudates.
17. The method of claim 15, wherein said sample is blood.
18. The method of claim 17, wherein said template DNA is obtained
from plasma from said blood.
19. The method of claim 17, wherein said template DNA is obtained
from serum from said blood.
20. The method of claim 15, wherein the recognition site is for a
Type IIS restriction enzyme.
21. The method of claim 20, wherein the Type IIS restriction enzyme
is selected from the group consisting of: Alw I, Alw26 I, Bbs I,
Bbv I, BceA I, Bmr I, Bsa I, Bst71 I, BsmA I, BsmB I, BsmF I, BspM
I, Ear I, Fau I, Fok I, Hga I, Pie I, Sap I, SSfaN I, and Sthi32 I.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of U.S. patent
application Ser. No. 10/661,165, filed Sep. 11, 2003, which is: (a)
a continuation-in-part of PCT/US03/06198, filed Feb. 28, 2003,
which claims benefit under 35 U.S.C. .sctn.119(e) of U.S.
Provisional Patent Application No. 60/378,354, filed May 8, 2002,
and which is a continuation-in-part of U.S. patent application Ser.
No. 10/093,618, filed Mar. 11, 2002, which claims benefit under 35
U.S.C. .sctn.119(e) of U.S. Provisional Patent Application No.
60/360,232, filed Mar. 1, 2002; (b) a continuation-in-part of
PCT/US03/27308, filed Aug. 29, 2003; and (c) a continuation-in-part
of U.S. patent application Ser. No. 10/376,770, filed Feb. 28,
2003, which claims benefit under 35 U.S.C. .sctn.119(e) of U.S.
Provisional Patent Application No. 60/378,354, filed May 8, 2002,
and which is a continuation-in-part of U.S. patent application Ser.
No. 10/093,618, filed Mar. 11, 2002, which claims benefit under 35
U.S.C. .sctn.119(e) of U.S. Provisional Patent Application No.
60/360,232, filed Mar. 1, 2002. The contents of these applications
are incorporated by reference in their entirety.
BACKGROUND OF THE INVENTION
[0002] 1. Field of the Invention
[0003] The present invention is directed to a method for the
detection of genetic disorders including chromosomal abnormalities
and mutations. The present invention provides a rapid, non-invasive
method for determining the sequence of DNA from a fetus. The method
is especially useful for detection of chromosomal abnormalities in
a fetus including translocations, transversions, monosomies,
trisomies, and other aneuploidies, deletions, additions,
amplifications, translocations and rearrangements.
[0004] 2. Background Art
[0005] Chromosomal abnormalities are responsible for a significant
portion of genetic defects in liveborn humans. The nucleus of a
human cell contains forty-six (46) chromosomes, which contain the
genetic instructions, and determine the operations of the cell.
Half of the forty-six chromosomes originate from each parent.
Except for the sex chromosomes, which are quite different from each
other in normal males, the chromosomes from the mother and the
chromosomes from the father make a matched set. The pairs were
combined when the egg was fertilized by the sperm. Occasionally, an
error occurs in either the formation or combination of chromosomes,
and the fertilized egg is formed with too many or too few
chromosomes, or with chromosomes that are mixed in some way.
Because each chromosome contains many genes, chromosomal
abnormalities are likely to cause serious birth defects, affecting
many body systems and often including developmental disability
(e.g., mental retardation).
[0006] Cells mistakenly can rejoin broken ends of chromosomes, both
spontaneously and after exposure to chemical compounds,
carcinogens, and irradiation. When rejoining occurs within a
chromosome, a chromosome segment between the two breakpoints
becomes inverted and is classified as an inversion. With
inversions, there is no loss of genetic material; however,
inversions can cause disruption of a critical gene, or create a
fusion gene that induces a disease related condition.
[0007] In a reciprocal translocation, two non-homologous
chromosomes break and exchange fragments. In this scenario, two
abnormal chromosomes result: each consists of a part derived from
the other chromosome and lacks a part of itself. If the
translocation is of a balanced type, the individual will display no
abnormal phenotypes. However, during germ-cell formation in the
translocation-bearing individuals, the proper distribution of
chromosomes in the egg or sperm occasionally fails, resulting in
miscarriage, malformation, or mental retardation of the
offspring.
[0008] In a Robertsonian translocation, the centromeres of two
acrocentric (a chromosome with a non-centrally located centromere)
chromosomes fuse to generate one large metacentric chromosome. The
karyotype of an individual with a centric fusion has one less than
the normal diploid number of chromosomes.
[0009] Errors that generate too many or too few chromosomes can
also lead to disease phenotypes. For example, a missing copy of
chromosome X (monosomy X) results in Turner's Syndrome, while an
additional copy of chromosome 21 results in Down's Syndrome. Other
diseases such as Edward's Syndrome, and Patau Syndrome are caused
by an additional copy of chromosome 18, and chromosome 13,
respectively.
[0010] One of the most common chromosome abnormalities is known as
Down syndrome. The estimated incidence of Down's syndrome is
between 1 in 1,000 to 1 in 1,100 live births. Each year
approximately 3,000 to 5,000 children are born in the U.S. with
this chromosomal disorder. The vast majority of children with Down
syndrome (approximately 95 percent) have an extra chromosome 21.
Most often, the extra chromosome originates from the mother.
However, in about 3-4 percent of people with Down syndrome, a
translocation between chromosome 21 and either 14 or 22 is
responsible for the genetic abnormality. Finally, another
chromosome problem, called mosaicism, is noted in about 1 percent
of individuals with Down's syndrome. In this case, some cells have
47 chromosomes and others have 46 chromosomes. Mosaicism is thought
to be the result of an error in cell division soon after
conception.
[0011] Chromosomal abnormalities are congenital, and therefore,
prenatal diagnosis can be used to determine the health and
condition of an unborn fetus. Without knowledge gained by prenatal
diagnosis, there could be an untoward outcome for the fetus or the
mother or both. Congenital anomalies account for 20 to 25% of
perinatal deaths. Specifically, prenatal diagnosis is helpful for
managing the remaining term of the pregnancy, planning for possible
complications with the birth process, preparing for problems that
can occur in the newborn infant, and finding conditions that may
affect future pregnancies.
[0012] There are a variety of non-invasive and invasive techniques
available for prenatal diagnosis including ultrasonography,
amniocentesis, chorionic villus sampling (CVS), fetal blood cells
in maternal blood, maternal serum alpha-fetoprotein, maternal serum
beta-HCG, and maternal serum estriol. However, the techniques that
are non-invasive are less specific, and the techniques with high
specificity and high sensitivity are highly invasive. Furthermore,
most techniques can be applied only during specific time periods
during pregnancy for greatest utility.
Ultrasonography
[0013] This is a harmless, non-invasive procedure. High frequency
sound waves are used to generate visible images from the pattern of
the echoes made by different tissues and organs, including the
fetus in the amniotic cavity. The developing embryo can be
visualized at about 6 weeks of gestation. The major internal organs
and extremities can be assessed to determine if any are abnormal at
about 16 to 20 weeks gestation.
[0014] An ultrasound examination can be useful to determine the
size and position of the fetus, the amount of amniotic fluid, and
the appearance of fetal anatomy; however, there are limitations to
this procedure. Subtle abnormalities, such as Down syndrome, where
the morphologic abnormalities are often not marked, but only
subtle, may not be detected at all.
Amniocentesis
[0015] This is a highly invasive procedure in which a needle is
passed through the mother's lower abdomen into the amniotic cavity
inside the uterus. This procedure can be performed at about 14
weeks gestation. For prenatal diagnosis, most amniocenteses are
performed between 14 and 20 weeks gestation. However, an ultrasound
examination is performed, prior to amniocentesis, to determine
gestational age, position of the fetus and placenta, and determine
if enough amniotic fluid is present. Within the amniotic fluid are
fetal cells (mostly derived from fetal skin) which can be grown in
culture for chromosomal, biochemical, and molecular biologic
analyses.
[0016] Large chromosomal abnormalities, such as extra or missing
chromosomes or chromosome fragments, can be detected by
karyotyping, which involves the identification and analysis of all
46 chromosomes from a cell and arranges them in their matched
pairs, based on subtle differences in size and structure. In this
systematic display, abnormalities in chromosome number and
structure are apparent. This procedure typically takes 7-10 days
for completion.
[0017] While amniocentesis can be used to provide direct genetic
information, risks are associated with the procedure including
fetal loss and maternal Rh sensitization. The increased risk for
fetal mortality following amniocentesis is about 0.5% above what
would normally be expected. Rh negative mothers can be treated with
RhoGam.
Chorionic Villus Sampling (CVS)
[0018] In this procedure, a catheter is passed via the vagina
through the cervix and into the uterus to the developing placenta
with ultrasound guidance. The introduction of the catheter allows
cells from the placental chorionic villi to be obtained and
analyzed by a variety of techniques, including chromosome analysis
to determine the karyotype of the fetus. The cells can also be
cultured for biochemical or molecular biologic analysis. Typically,
CVS is performed between 9.5 and 12.5 weeks gestation.
[0019] CVS has the disadvantage of being an invasive procedure, and
it has a low but significant rate of morbidity for the fetus; this
loss rate is about 0.5 to 1% higher than for women undergoing
amniocentesis. Rarely, CVS can be associated with limb defects in
the fetus. Also, the possibility of maternal Rh sensitization is
present. Furthermore, there is also the possibility that maternal
blood cells in the developing placenta will be sampled instead of
fetal cells and confound chromosome analysis.
Maternal Serum Alpha-Fetoprotein (MSAFP)
[0020] The developing fetus has two major blood proteins--albumin
and alpha-fetoprotein (AFP). The mother typically has only albumin
in her blood, and thus, the MSAFP test can be utilized to determine
the levels of AFP from the fetus. Ordinarily, only a small amount
of AFP gains access to the amniotic fluid and crosses the placenta
to mother's blood. However, if the fetus has a neural tube defect,
then more AFP escapes into the amniotic fluid. Neural tube defects
include anencephaly (failure of closure at the cranial end of the
neural tube) and spina bifida (failure of closure at the caudal end
of the neural tube). The incidence of such defects is about 1 to 2
births per 1000 in the United States. Also, if there are defects in
the abdominal wall, the AFP from the fetus will end up in maternal
blood in higher amounts.
[0021] The amount of MSAFP increases with gestational age, and thus
for the MSAFP test to provide accurate results, the gestational age
must be known with certainty. Also, the race of the mother and
presence of gestational diabetes can influence the level of MSAFP
that is to be considered normal. The MSAFP is typically reported as
multiples of the mean (MoM). The greater the MoM, the more likely a
defect is present. The MSAFP test has the greatest sensitivity
between 16 and 18 weeks gestation, but can be used between 15 and
22 weeks gestation. The MSAFP tends to be lower when Down's
Syndrome or other chromosomal abnormalities is present.
[0022] While the MSAFP test is non-invasive, the MSAFP is not 100%
specific. MSAFP can be elevated for a variety of reasons that are
not related to fetal neural tube or abdominal wall defects. The
most common cause for an elevated MSAFP is a wrong estimation of
the gestational age of the fetus. Therefore, results from an MSAFP
test are never considered definitive and conclusive.
Maternal Serum Beta-HCG
[0023] Beginning at about a week following conception and
implantation of the developing embryo into the uterus, the
trophoblast will produce detectable beta-HCG (the beta subunit of
human chorionic gonadotropin), which can be used to diagnose
pregnancy. The beta-HCG also can be quantified in maternal serum,
and this can be useful early in pregnancy when threatened abortion
or ectopic pregnancy is suspected, because the amount of beta-HCG
will be lower than normal.
[0024] In the middle to late second trimester, the beta-HCG can be
used in conjunction with the MSAFP to screen for chromosomal
abnormalities, in particular for Down syndrome. An elevated
beta-HCG coupled with a decreased MSAFP suggests Down syndrome.
High levels of HCG suggest trophoblastic disease (molar pregnancy).
The absence of a fetus on ultrasonography along with an elevated
HCG suggests a hydatidiform mole.
Maternal Serum Estriol
[0025] The amount of estriol in maternal serum is dependent upon a
viable fetus, a properly functioning placenta, and maternal
well-being. Dehydroepiandrosterone (DHEA) is made by the fetal
adrenal glands, and is metabolized in the placenta to estriol. The
estriol enters the maternal circulation and is excreted by the
maternal kidney in urine or by the maternal liver in the bile.
Normal levels of estriol, measured in the third trimester, will
give an indication of general well-being of the fetus. If the
estriol level drops, then the fetus is threatened and an immediate
delivery may be necessary. Estriol tends to be lower when Down
syndrome is present and when there is adrenal hypoplasia with
anencephaly.
The Triple Screen Test
[0026] The triple screen test comprises analysis of maternal serum
alpha-feto-protein (MSAFP), human chorionic gonadotrophin (hCG),
and unconjugated estriol (uE3). The blood test is usually performed
16-18 weeks after the last menstrual period. While the triple
screen test is non-invasive, abnormal test results are not
indicative of a birth defect. Rather, the test only indicates an
increased risk and suggests that further testing is needed. For
example, 100 out of 1,000 women will have an abnormal result from
the triple screen test. However, only 2-3 of the 100 women will
have a fetus with a birth defect. This high incidence of false
positives causes tremendous stress and unnecessary anxiety to the
expectant mother.
Fetal Cells Isolated from Maternal Blood
[0027] The presence of fetal nucleated cells in maternal blood
makes it possible to use these cells or noninvasive prenatal
diagnosis (Walknowska, et al., Lancet 1:1119-1122, 1969; Lo et al.,
Lancet 2:1363-65, 1989; Lo et al., Blood 88:4390-95, 1996). The
fetal cells can be sorted and analyzed by a variety of techniques
to look for particular DNA sequences (Bianchi et al., Am. J. Hum.
Genet. 61:822-29, (1997); Bianchi et al., PNAS 93:705-08, (1996)).
Fluorescence in-situ hybridization (FISH) is one technique that can
be applied to identify particular chromosomes of the fetal cells
recovered from maternal blood and diagnose aneuploid conditions
such as trisomies and monosomy X. Also, it has been reported that
the number of fetal cells in maternal blood increases in aneuploid
pregnancies.
[0028] The method of FISH uses DNA probes labeled with colored
fluorescent tags that allow detection of specific chromosomes or
genes under a microscope. Using FISH, subtle genetic abnormalities
that cannot be detected by standard karyotyping are readily
identifiable. This procedure typically takes 24-48 hours to
complete. Additionally, using a panel of multi-colored DNA FISH
probes, abnormal chromosome copy numbers can be seen.
[0029] While improvements have been made for the isolation and
enrichment of fetal cells, it is still difficult to get many fetal
blood cells. There may not be enough to reliably determine
anomalies of the fetal karyotype or assay for other abnormalities.
Furthermore, most techniques are time consuming, require
high-inputs of labor, and are difficult to implement for a high
throughput fashion.
Fetal DNA from Maternal Blood
[0030] Fetal DNA has been detected and quantitated in maternal
plasma and serum (Lo et al., Lancet 350:485-487 (1997); Lo et al.,
Am. J. hum. Genet. 62:768-775 (1998)). Multiple fetal cell types
occur in the maternal circulation, including fetal granulocytes,
lymphocytes, nucleated red blood cells, and trophoblast cells
(Pertl, and Bianchi, Obstetrics and Gynecology 98: 483-490 (2001)).
Fetal DNA can be detected in the serum at the seventh week of
gestation, and increases with the term of the pregnancy. The fetal
DNA present in the maternal serum and plasma is comparable to the
concentration of DNA obtained from fetal cell isolation
protocols.
[0031] Circulating fetal DNA has been used to determine the sex of
the fetus (Lo et al., Am. J. hum. Genet. 62:768-775 (1998)). Also,
fetal rhesus D genotype has been detected using fetal DNA. However,
the diagnostic and clinical applications of circulating fetal DNA
is limited to genes that are present in the fetus but not in the
mother (Pertl and Bianchi, Obstetrics and Gynecology 98: 483-490
(2001)). Thus, a need still exists for a non-invasive method that
can determine the sequence of fetal DNA and provide definitive
diagnosis of chromosomal abnormalities in a fetus.
BRIEF SUMMARY OF THE INVENTION
[0032] The invention is directed to a method for detection of
genetic disorders including mutations and chromosomal
abnormalities. In some embodiments, the present invention is used
to detect mutations, and chromosomal abnormalities including but
not limited to translocation, transversion, monosomy, trisomy, and
other aneuploidies, deletion, addition, amplification, fragment,
translocation, and rearrangement. Numerous abnormalities can be
detected simultaneously. The present invention also provides a
non-invasive method to determine the sequence of fetal DNA from a
sample of a pregnant female. The present invention can be used to
detect any alternation in gene sequence as compared to the wild
type sequence including but not limited to point mutation, reading
frame shift, transition, transversion, addition, insertion,
deletion, addition-deletion, frame-shift, missense, reverse
mutation, and microsatellite alteration. The present invention also
provides a method for isolating free nucleic acid from a sample
containing nucleic acid. The present invention also provides
compositions and kits.
[0033] In one aspect, the invention is directed to methods for
detecting chromosomal abnormalities. In one embodiment, the present
invention is directed to a method for detecting chromosomal
abnormalities, said method comprising quantitating the relative
amount of the alleles at a heterozygous locus of interest, where
the heterozygous locus of interest was previously identified by
determining the sequence of alleles at a locus of interest from
template DNA, wherein said relative amount is expressed as a ratio,
and wherein said ratio indicates the presence or absence of a
chromosomal abnormality.
[0034] In some embodiments, determining the sequence includes using
a method that is allele specific PCR, mass spectrometry,
hybridization, primer extension, fluorescence resonance energy
transfer (FRET), sequencing, Sanger dideoxy sequencing, DNA
microarray, GeneCHIP arrays, HuSNP arrays, CodeLink Arrays,
BeadArray Technology, MassARRAY, MassEXTEND, SNP-IT, TaqMan,
InvaderStrand Assay, southern blot, slot blot, dot blot, or
MALDI-TOF mass spectrometry.
[0035] In some embodiments, template DNA is obtained from human,
non-human, mammal, reptile, cattle, cat, dog, goat, swine, pig,
monkey, ape, gorilla, bull, cow, bear, horse, sheep, poultry,
mouse, rat, fish, dolphin, whale, or shark. In an embodiment, the
template DNA is obtained from a human source. In a preferred
embodiment, the template DNA is obtained from a pregnant human
female. In some embodiments, the template DNA is obtained from a
sample that is a cell, fetal cell, tissue, blood, serum, plasma,
saliva, urine, tear, vaginal secretion, sweat, umbilical cord
blood, chorionic villi amniotic fluid, embryonic tissue, an embryo,
a two-celled embryo, a four-celled embryo, an eight celled embryo,
a 16-celled embryo, a 32-celled embryo, a 64-celled embryo, a
128-celled embryo, a 256-celled embryo, a 512-celled embryo, a
1024-celled embryo, lymph fluid, cerebrospinal fluid, mucosa
secretion, peritoneal fluid, ascitic fluid, fecal matter, or body
exudates. In these embodiments, the sample may be mixed with an
agent that inhibits cell lysis to inhibit the lysis of cells, if
cells are present, where the agent is a membrane stabilizer, a
cross-linker, or a cell lysis inhibitor. In some of these
embodiments, agent is a cell lysis inhibitor, and may be
glutaraldehyde, derivatives of glutaraldehyde, formaldehyde,
formalin, or derivatives of formaldehyde. In some of these
embodiments the sample is blood and in one embodiment the sample is
blood from a pregnant female, e.g., a human female. In the latter
embodiment, the fetus may be at a gestational age selected from the
group consisting of: 0-4, 4-8, 8-12, 12-16, 16-20, 20-24, 24-28,
28-32, 32-36, 36-40, 40-44, 44-48, 48-52, or more than 52 weeks. In
some of these embodiments, the template DNA may be obtained from
plasma or from serum from the blood. In these embodiments, the
template DNA may include a mixture of maternal DNA and fetal DNA,
and in one embodiment, prior to determining the sequence of alleles
of a locus of interest from template DNA, maternal DNA is sequenced
to identify a homozygous locus of interest, and the homozygous
locus of interest is the locus of interest analyzed in the template
DNA. In another embodiment, maternal DNA is sequenced to identify a
heterozygous locus of interest, and the heterozygous locus of
interest is the locus of interest analyzed in the template DNA.
[0036] In embodiments, alleles of multiple loci of interest are
sequenced and their relative amounts quantitated and expressed as a
ratio. In one embodiment, the sequence of alleles of one to tens to
hundreds to thousands of loci of interest on a single chromosome on
template DNA is determined. In another embodiment, the sequence of
alleles of one to tens to hundreds to thousands of loci of interest
on multiple chromosomes is determined.
[0037] In an embodiment, the locus of interest is suspected of
containing a single nucleotide polymorphism or mutation. The method
can be used for determining sequences of multiple loci of interest
concurrently. The template DNA can comprise multiple loci from a
single chromosome. The template DNA can comprise multiple loci from
different chromosomes. The loci of interest on template DNA can be
amplified in one reaction. Alternatively, each of the loci of
interest on template DNA can be amplified in a separate reaction.
The amplified DNA can be pooled together prior to digestion of the
amplified DNA. Each of the labeled DNA containing a locus of
interest can be separated prior to determining the sequence of the
locus of interest. In one embodiment, at least one of the loci of
interest is suspected of containing a single nucleotide
polymorphism or a mutation.
[0038] There is no limitation as to the chromosomes that can be
compared. The ratio for the alleles at a heterozygous locus of
interest on any chromosome can be compared to the ratio for the
alleles at a heterozygous locus of interest on any other
chromosome. In another embodiment, the ratio of alleles at a
heterozygous locus of interest on a chromosome is compared to the
ratio of alleles at a heterozygous locus of interest on two, three,
four or more than four chromosomes. In another embodiment, the
ratio of alleles at multiple loci of interest on a chromosome is
compared to the ratio of alleles at multiple loci of interest on
two, three, four, or more than four chromosomes. In embodiments,
the ratio for alleles at heterozygous loci of interest on a
chromosome are summed and compared to the ratio for alleles at
heterozygous loci of interest on a different chromosome, where a
difference in ratios indicates the presence of a chromosomal
abnormality. In some of these embodiments, the chromosomes that are
compared are human chromosomes such as chromosome 1, 2, 3, 4, 5, 6,
7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X, or
Y. In one of the latter embodiments, the ratio for the alleles at
heterozygous loci of interest of chromosomes 13, 18, and 21 are
compared. In another embodiment, the sequence of one to tens to
hundreds to thousands of loci of interest on the template DNA
obtained from a sample of a pregnant female is determined. In one
embodiment, the loci of interest are on one chromosome. In another
embodiment, the loci of interest are on multiple chromosomes.
[0039] In some embodiments, determining the sequence of the alleles
comprises amplifying alleles of a locus of interest on a template
DNA using a first and a second primer, where the second primer
contains a recognition site for a restriction enzyme such that
digestion with the restriction enzyme generates a 5' overhang
containing the locus of interest; digesting the amplified DNA with
the restriction enzyme that recognizes the recognition site on the
second primer; incorporating a nucleotide into the digested DNA by
using the 5' overhang containing the locus of interest as a
template; and determining the sequence of the alleles of the locus
of interest by determining the sequence of the DNA into which the
nucleotide was incorporated. In one embodiment, determination of
the sequence of the locus of interest in comprises detecting a
nucleotide.
[0040] In other embodiments, determining the sequence of alleles
comprises amplifying alleles of a locus of interest on a template
DNA using a first and second primers, where the second primer
contains a recognition site for a restriction enzyme such that
digestion with the restriction enzyme generates a 5' overhang
containing the locus of interest; digesting the amplified DNA with
the restriction enzyme that recognizes the recognition site on the
second primer; incorporating nucleotides into the digested DNA of
(b), where a nucleotide that terminates elongation, and is
complementary to the locus of interest of an allele, is
incorporated into the 5' overhang of said allele, and a nucleotide
complementary to the locus of interest of a different allele is
incorporated into the 5' overhang of said different allele, and
said terminating nucleotide, which is complementary to a nucleotide
in the 5' overhang of said different allele, is incorporated into
the 5' overhang of said different allele; and determining the
sequence of the alleles of a locus of interest by determining the
sequence of the DNA into which the complementary nucleotides have
been incorporated. In one embodiment, determination of the sequence
of the locus of interest comprises detecting a nucleotide.
[0041] The incorporation of a nucleotide may be accomplished by a
DNA polymerase, including but not limited to E. coli DNA
polymerase, Klenow fragment of E. coli DNA polymerase I, T7 DNA
polymerase, T4 DNA polymerase, T5 DNA polymerase, Klenow class
polymerases, Taq polymerase, bacteriophage 29, REDTaq.TM. Genomic
DNA polymerase, Pfu DNA polymerase, Vent DNA polymerase or
sequenase. Incorporation of a nucleotide may include incorporation
of a labeled nucleotide, or labeled and unlabeled nucleotides. One
nucleotide, two nucleotides, three nucleotides, four nucleotides,
five nucleotides, or more than five nucleotides can be
incorporated. A combination of labeled and unlabeled nucleotides
can be incorporated. The labeled nucleotide may be a
dideoxynucleotide triphosphate (also referred to as "dideoxy") or
deoxynucleotide triphosphate (also referred to as "deoxy"). The
unlabeled nucleotide may be a dideoxynucleotide triphosphate or
deoxynucleotide triphosphate. Labeled nucleotides may be labeled
with a radioactive molecule, fluorescent molecule, antibody,
antibody fragment, hapten, carbohydrate, biotin, derivative of
biotin, phosphorescent moiety, luminescent moiety,
electrochemiluminescent moiety, chromatic moiety, and moiety having
a detectable electron spin resonance, electrical capacitance,
dielectric constant or electrical conductivity. In one embodiment,
the labeled nucleotide is labeled with a fluorescent molecule. The
incorporation of a fluorescent labeled nucleotide may further
comprise using a mixture of fluorescent and unlabeled
nucleotides.
[0042] In one embodiment, the determination of the sequence of the
locus of interest comprises detecting the incorporated nucleotide.
The detection method includes but is not limited to gel
electrophoresis, capillary electrophoresis, microchannel
electrophoresis, polyacrylamide gel electrophoresis, fluorescence
detection, fluorescence polarization, DNA sequencing, Sanger
dideoxy sequencing, ELISA, mass spectrometry, time of flight mass
spectrometry, quadrupole mass spectrometry, magnetic sector mass
spectrometry, electric sector mass spectrometry, fluorometry,
infrared spectrometry, ultraviolet spectrometry, palentiostatic
amperometry, DNA hybridization, DNA microarray, GeneChip arrays,
HuSNP arrays, BeadArrays, MassExtend, SNP-IT, TaqMan assay, Invader
assay, MassCleave, southern blot, slot blot, or dot blot.
[0043] In embodiments, first and second primers contain a portion
of a restriction enzyme recognition site that contains a variable
nucleotide, where the full restriction enzyme recognition site is
generated after amplification. In some embodiments, the 3' region
of said primers can contain mismatches with the template DNA, and
digestion with said restriction enzyme generates a 5' overhang
containing the locus of interest. In some embodiments, the
restriction enzyme recognition site is for a restriction enzyme
that includes but is not limited to BsaJ I, Bssk I, Dde I, EcoN I,
Fnu4H I, Hinf I, or ScrF I. In some embodiments, the restriction
enzyme cuts DNA at a distance, from the recognition site. In some
of these embodiments, the recognition site is for a Type IIS
restriction enzyme. In some of these embodiments, the Type IIS
restriction enzyme includes but is not limited to Alw I, Alw26 I,
Bbs I, Bbv I, BceA I, Bmr I, Bsa I, Bst71 I, BsmA I, BsmB I, BsmF
I, BspM I, Ear I, Fau I, Fok I, Hga I, Ple I, Sap I, SSfaN I, or
Sthi32 I.
[0044] In some embodiments, the recognition site for restriction
enzymes includes but is not limited to BsaJ I (5'
C.sup..dwnarw.CNNGG 3'), BssK I (5'.sup..dwnarw.CCNGG 3'), Dde I
(5'C.sup..dwnarw.TNAG 3'), EcoN I (5'CCTNN.sup..dwnarw.NNNAGG 3'
(SEQ ID NO: 7)), Fnu4H I (5'GC.sup..dwnarw.NGC 3'), Hinf I
(5'G.sup..dwnarw.ANTC 3'), PflF I (5' GACN.sup..dwnarw.NNGTC 3'),
Sau96 I (5' G.sup..dwnarw.GNCC 3'), ScrF I (5' CC.sup..dwnarw.NGG
3'), Tth1 11 I (5' GACN.sup..dwnarw.NNGTC 3'), and more preferably
Fnu4H I and EcoN I, is generated after amplification.
[0045] The first and/or second primer can contain a tag at the 5'
terminus. In some embodiments, the first primer contains a tag at
the 5' terminus. The tag can be used to separate the amplified DNA
from the template DNA. The tag can be used to separate the
amplified DNA containing the labeled nucleotide from the amplified
DNA that does not contain the labeled nucleotide. The tag can be
any chemical moiety including but not limited to radioisotope,
fluorescent reporter molecule, chemiluminescent reporter molecule,
antibody, antibody fragment, hapten, biotin, derivative of biotin,
photobiotin, iminobiotin, digoxigenin, avidin, enzyme, acridinium,
sugar, enzyme, apoenzyme, homopolymeric oligonucleotide, hormone,
ferromagnetic moiety, paramagnetic moiety, diamagnetic moiety,
phosphorescent moiety, luminescent moiety, electrochemiluminescent
moiety, chromatic moiety, moiety having a detectable electron spin
resonance, electrical capacitance, dielectric constant or
electrical conductivity, or combinations thereof. In some
embodiments, the tag is biotin. The biotin tag is used to separate
amplified DNA from the template DNA using a streptavidin matrix.
The streptavidin matrix may be coated on wells of a microtiter
plate.
[0046] In some embodiments, the annealing length of the second
primer is selected from the group consisting of 35-30, 30-25,
25-20, 20-15, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, and less
than 4 bases.
[0047] In embodiments, the method of amplification includes but is
not limited to polymerase chain reaction, self-sustained sequence
reaction, ligase chain reaction, rapid amplification of cDNA ends,
polymerase chain reaction and ligase chain reaction, Q-beta phage
amplification, strand displacement amplification, or splice overlap
extension polymerase chain reaction. In some embodiments, the
method of amplification is PCR. In some of these embodiments, an
annealing temperature for cycle 1 of PCR is about the melting
temperature of the portion of the 3' region of the second primer
that anneals to the template DNA. In some of the latter
embodiments, an annealing temperature for cycle 2 of PCR is about
the melting temperature of the portion of the 3' region of the
first primer that anneals to the template DNA. In some of the
latter embodiments, an annealing temperature for the remaining
cycles of PCR is at about the melting temperature of the entire
second primer.
[0048] In another aspect, the invention provides methods of
determining the sequence of a locus of interest from a sample
comprising free fetal DNA, where an agent that inhibits cell lysis
has been added to the sample to inhibit lysis of cells, if cells
are present, where the agent is a membrane stabilizer, a
cross-linker, or a cell lysis inhibitor.
[0049] In some embodiments, the agent is a cell lysis inhibitor,
and in some of these embodiments, the cell lysis inhibitor includes
but is not limited to glutaraldehyde, derivatives of
glutaraldehyde, formaldehyde, derivatives of formaldehyde, or
formalin. In embodiments, the sample includes but is not limited to
tissue, cell, blood, serum, plasma, urine, or vaginal secretion. In
some embodiments, the sample is blood. In some of these
embodiments, the template DNA is isolated from the serum, in other
embodiments the template DNA is isolated from plasma. In some
embodiments, the sample contains free maternal template DNA and
free fetal template DNA. In some embodiments, prior to determining
the sequence, template DNA was isolated. In some embodiments, prior
to determining the sequence of the locus of interest on fetal DNA,
the sequence of the locus of interest on maternal template DNA was
determined. In some embodiments, prior to determining the sequence
of the locus of interest on fetal DNA, the sequence of the locus of
interest on paternal template DNA was determined. In some
embodiments, the locus of interest is a single nucleotide
polymorphism. In other embodiments, the locus of interest is a
mutation. In some embodiments, the sequence of multiple loci of
interest is determined. In some of these embodiments, the multiple
loci of interest are on multiple chromosomes.
[0050] In some embodiments, the sequence is determined by: (a)
amplifying a locus of interest on a template DNA using a first and
second primers, where the second primer contains a recognition site
for a restriction enzyme such that digestion with the restriction
enzyme generates a 5' overhang containing the locus of interest;
(b) digesting the amplified DNA with the restriction enzyme that
recognizes the recognition site on the second primer; (c)
incorporating a nucleotide into the digested DNA of (b) by using
the 5' overhang containing the locus of interest as a template; and
(d) determining the sequence of the locus of interest by
determining the sequence of the DNA of (c).
[0051] In other embodiments, the sequence is determined by: (a)
amplifying alleles of a locus of interest on a template DNA using a
first and second primers, where the second primer contains a
recognition site for a restriction enzyme such that digestion with
the restriction enzyme generates a 5' overhang containing the locus
of interest; (b) digesting the amplified DNA with the restriction
enzyme that recognizes the recognition site on the second primer;
(c) incorporating nucleotides into the digested DNA of (b), where a
nucleotide that terminates elongation, and is complementary to the
locus of interest of an allele, is incorporated into the 5'
overhang of said allele, and a nucleotide complementary to the
locus of interest of a different allele is incorporated into the 5'
overhang of said different allele, and the terminating nucleotide,
which is complementary to a nucleotide in the 5' overhang of said
different allele, is incorporated into the 5' overhang of said
different allele; and (d) determining the sequence of the alleles
of a locus of interest by determining the sequence of the DNA of
(c).
[0052] In some embodiments, the restriction enzyme cuts DNA at a
distance from the recognition site. In some of these embodiments,
the recognition site includes but is for a Type IIS restriction
enzyme, for example Alw I, Alw26 I, Bbs I, Bbv I, BceA I, Bmr I,
Bsa I, Bst71 I, BsmA I, BsmB I, BsmF I, BspM I, Ear I, Fau I, Fok
I, Hga I, Ple I, Sap I, SSfaN I, or Sthi32 I.
[0053] In some embodiments, the method of amplification may be, for
example, polymerase chain reaction, self-sustained sequence
reaction, ligase chain reaction, rapid amplification of cDNA ends,
polymerase chain reaction and ligase chain reaction, Q-beta phage
amplification, strand displacement amplification, or splice overlap
extension polymerase chain reaction. In some embodiments, the
method of amplification is by PCR. In some of these embodiments, an
annealing temperature for cycle 1 of PCR is about the melting
temperature of the portion of the 3' region of the second primer
that anneals to the template DNA. In some of the latter
embodiments, an annealing temperature for cycle 2 of PCR is about
the melting temperature of the portion of the 3' region of the
first primer that anneals to the template DNA. In some of the
latter embodiments, an annealing temperature for the remaining
cycles of PCR is at about the melting temperature of the entire
second primer.
[0054] In some embodiments, the sequence of a locus of interest was
determined using allele specific PCR, mass spectrometry,
hybridization, primer extension, fluorescence polarization,
fluorescence resonance energy transfer (FRET), fluorescence
detection, sequencing, Sanger dideoxy sequencing, DNA microarray,
southern blot, slot blot, dot blot, or MALDI-TOF mass
spectrometry.
[0055] In some embodiments, the sequence of a locus of interest is
determined by (1) amplification of the locus of interest; (2)
hybridization of amplified loci to GeneCHIP array (3) washing
GeneCHIP array; (4) staining the GeneCHIP array with detectable
reagents; and (5) scanning GeneCHIP array. In some of these
embodiments, the amplification method in (1) is polymerase chain
reaction, self-sustained sequence reaction, ligase chain reaction,
rapid amplification of cDNA ends, polymerase chain reaction and
ligase chain reaction, Q-beta phage amplification, strand
displacement amplification, or splice overlap extension polymerase
chain reaction. In some embodiments, the method of amplification is
by PCR. In some embodiments, the staining method comprises
streptavidin phycoerythrin and biotinylated anti-streptavidin. In
some embodiments, an agent that inhibits cell lysis has been added
to the sample to inhibit the lysis of cells, if present, where the
agent is membrane stabilizer, cross-linker, or cell lysis
inhibitor. In some embodiments, the agent is a cell lysis
inhibitor. In some of these embodiments, the cell lysis inhibitor
is formalin at a percentage selected from the group consisting of:
0.0001-0.03%, 0.03-0.05%, 0.05-0.08%, 0.08-0.1%, 0.1-0.3%,
0.3-0.5%, 0.5-0.7%, 0.7-0.9%, 0.9-1.2%, 1.2-1.5%, 1.5-2%, or 2-3%.
In some embodiments, the concentration of formalin in the sample is
0.1%.
[0056] In some embodiments, the sequence of a locus of interest is
determined by (1) amplification of the locus of interest; (2)
amplicon fragmentation; (3) hybridization of fragmented amplicons
to CodeLink Arrays; (4) extension reaction to incorporate a
nucleotide; and (5) detection of incorporated nucleotides. In some
of these embodiments, the amplification method is polymerase chain
reaction, self-sustained sequence reaction, ligase chain reaction,
rapid amplification of cDNA ends, polymerase chain reaction and
ligase chain reaction, Q-beta phage amplification, strand
displacement amplification, or splice overlap extension polymerase
chain reaction. In some embodiments, the method of amplification is
by PCR. In some embodiments, the amplicon fragmentation is by
exonuclease digestion. In some embodiments, the incorporated
nucleotide is a dideoxynucleotide or deoxynucleotide. In some
embodiments, the incorporated nucleotide is labeled with a
radioactive molecule, fluorescent molecule, antibody, antibody
fragment, hapten, carbohydrate, biotin, derivative of biotin,
phosphorescent moiety, luminescent moiety, electrochemiluminescent
moiety, chromatic moiety, and moiety having a detectable electron
spin resonance, electrical capacitance, dielectric constant or
electrical conductivity. In some embodiments, the labeled
nucleotide is labeled with a fluorescent molecule. In some
embodiments, an agent that inhibits cell lysis has been added to
the sample to inhibit the lysis of cells, if present, where the
agent is membrane stabilizer, cross-linker, or cell lysis
inhibitor. In some embodiments, the agent is a cell lysis
inhibitor. In some of these embodiments, the cell lysis inhibitor
is formalin at a percentage selected from the group consisting of:
0.0001-0.03%, 0.03-0.05%, 0.05-0.08%, 0.08-0.1%, 0.1-0.3%,
0.3-0.5%, 0.5-0.7%, 0.7-0.9%, 0.9-1.2%, 1.5-2%, or 2-3%. In some
embodiments, the concentration of formalin in the sample is
0.1%.
[0057] In some embodiments, the sequence of a locus of interest is
determined by using BeadArray Technology. In some embodiments, an
agent that inhibits cell lysis has been added to the sample to
inhibit the lysis of cells, if present, where the agent is membrane
stabilizer, cross-linker, or cell lysis inhibitor. In some
embodiments, the agent is a cell lysis inhibitor. In some of these
embodiments, the cell lysis inhibitor is formalin at a percentage
selected from the group consisting of: 0.0001-0.03%, 0.03-0.05%,
0.05-0.08%, 0.08-0.1%, 0.1-0.3%, 0.3-0.5%, 0.5-0.7%, 0.7-0.9%,
0.9-1.2%, 1.2-1.5%, 1.5-2%, or 2-3%. In some embodiments, the
concentration of formalin in the sample is 0.1%.
[0058] In some embodiments, the sequence of a locus of interest is
determined by (1) amplification of the locus of interest; (2)
dephosphorylation of the unused reagents in (1); (3) in vitro
transcription reaction of the products of (2); (4) RNase A cleavage
of the products of (3); (5) mixing the products of (4) with
CleanResin; (6) transfer products of (5) to SpectroCHIP; and (7)
analysis of the SpectroCHIP. In some of these embodiments, the
amplification method is polymerase chain reaction, self-sustained
sequence reaction, ligase chain reaction, rapid amplification of
cDNA ends, polymerase chain reaction and ligase chain reaction,
Q-beta phage amplification, strand displacement amplification, or
splice overlap extension polymerase chain reaction. In some
embodiments, the method of amplification is by PCR. In some
embodiments, the dephosphorylation reaction is catalyzed by shrimp
alkaline phosphatase. In some embodiments, an agent that inhibits
cell lysis has been added to the sample to inhibit the lysis of
cells, if present, where the agent is membrane stabilizer,
cross-linker, or cell lysis inhibitor. In some embodiments, the
agent is a cell lysis inhibitor. In some of these embodiments, the
cell lysis inhibitor is formalin at a percentage selected from the
group consisting of 0.0001-0.03%, 0.03-0.05%, 0.05-0.08%,
0.08-0.1%, 0.1-0.3%, 0.3-0.5%, 0.5-0.7%, 0.7-0.9%, 0.9-1.2%,
1.2-1.5%, 1.5-2%, or 2-3%. In some embodiments, the concentration
of formalin in the sample is 0.1%.
[0059] In some embodiments, the sequence of a locus of interest is
determined by (1) amplification of a locus of interest; (2)
dephosphorylation of the unused reagents in (1); (3) hybridization
of a primer to the locus of interest; (4) incorporation of a
nucleotide; (5) mixing the products of (4) with CleanResin; (6)
transfer products of (5) to SpectroCHIP; and (7) analysis of the
SpectroCHIP. In some of these embodiments, the amplification method
is polymerase chain reaction, self-sustained sequence reaction,
ligase chain reaction, rapid amplification of cDNA ends, polymerase
chain reaction and ligase chain reaction, Q-beta phage
amplification, strand displacement amplification, or splice overlap
extension polymerase chain reaction. In some embodiments, the
method of amplification is by PCR. In some embodiments, the
dephosphorylation reaction is catalyzed by shrimp alkaline
phosphatase. In some embodiments, hybridization of primer is
adjacent to the locus of interest. In some embodiments, the
incorporated nucleotide is a dideoxynucleotide or deoxynucleotide.
In some embodiments, the incorporated nucleotide is labeled with
radioactive molecule, fluorescent molecule, antibody, antibody
fragment, hapten, carbohydrate, biotin, derivative of biotin,
phosphorescent moiety, luminescent moiety, electrochemiluminescent
moiety, chromatic moiety, and moiety having a detectable electron
spin resonance, electrical capacitance, dielectric constant or
electrical conductivity. In some embodiments, the labeled
nucleotide is labeled with a fluorescent molecule. In some
embodiments, an agent that inhibits cell lysis has been added to
the sample to inhibit the lysis of cells, if present, where the
agent is membrane stabilizer, cross-linker, or cell lysis
inhibitor. In some embodiments, the agent is a cell lysis
inhibitor. In some of these embodiments, the cell lysis inhibitor
is formalin at a percentage selected from the group consisting of:
0.0001-0.03%, 0.03-0.05%, 0.05-0.08%, 0.08-0.1%, 0.1-0.3%,
0.3-0.5%, 0.5-0.7%, 0.7-0.9%, 0.9-1.2%, 1.2-1.5%, 1.5-2%, or 2-3%.
In some embodiments, the concentration of formalin in the sample is
0.1%.
[0060] In some embodiments, the sequence of a locus of interest is
determined by (1) amplification of the locus of interest; (2)
exonuclease treatment of the products of (1); (3) single stranded
DNA of (2) is annealed to an oligonucleotide; (4) incorporation of
a nucleotide using the annealed template and primer of (3); (5)
detection of the incorporated nucleotide. In some embodiments, the
amplification method is by polymerase chain reaction,
self-sustained sequence reaction, ligase chain reaction, rapid
amplification of cDNA ends, polymerase chain reaction and ligase
chain reaction, Q-beta phage amplification, strand displacement
amplification, or splice overlap extension polymerase chain
reaction. In some embodiments, the method of amplification is by
PCR. In some embodiments, the primer hybridizes adjacent to the
locus of interest. In some embodiment, the incorporated nucleotide
is a dideoxynucleotide or deoxynucleotide. In some embodiments, the
incorporation reaction comprises two terminating nucleotides and
two non-terminating nucleotides. In some embodiments, the
incorporated nucleotide is labeled with radioactive molecule,
fluorescent molecule, antibody, antibody fragment, hapten,
carbohydrate, biotin, derivative of biotin, phosphorescent moiety,
luminescent moiety, electrochemiluminescent moiety, chromatic
moiety, and moiety having a detectable electron spin resonance,
electrical capacitance, dielectric constant or electrical
conductivity. Income embodiments, the terminating nucleotides are
labeled with radioactive molecule, fluorescent molecule, antibody,
antibody fragment, hapten, carbohydrate, biotin, derivative of
biotin, phosphorescent moiety, luminescent moiety,
electrochemiluminescent moiety, chromatic moiety, and moiety having
a detectable electron spin resonance, electrical capacitance,
dielectric constant or electrical conductivity. In some
embodiments, the labeled nucleotide is labeled with a fluorescent
molecule. In some embodiments, the terminating nucleotides are
labeled with a fluorescent molecule. In some embodiments, an agent
that inhibits cell lysis has been added to the sample to inhibit
the lysis of cells, if present, where the agent is membrane
stabilizer, cross-linker, or cell lysis inhibitor. In some
embodiments, the agent is a cell lysis inhibitor. In some of these
embodiments, the cell lysis inhibitor is formalin at a percentage
selected from the group consisting of: 0.0001-0.03%, 0.03-0.05%,
0.05-0.08%, 0.08-0.1%, 0.1-0.3%, 0.3-0.5%, 0.5-0.7%, 0.7-0.9%,
09-1.2%, 1.2-1.5%, 1.5-2%, or 2-3%. In some embodiments, the
concentration of formalin in the sample is 0.1%.
[0061] In some embodiments, the sequence of a locus of interest is
determined by (1) amplification of the locus of interest, wherein
the amplification reaction comprises a forward primer, a reverse
primer, and a probe that anneals to the locus of interest, which is
within the region of the amplicon; and (2) detection of the PCR
products, wherein the amount of PCR product is used to determine
the presence or absence of a specific genetic sequence. In some
embodiments, the amplification is by PCR. In some embodiments, the
probe contains a reporter dye at the 5' end and the 3' end contains
a quenching dye. In some embodiments, the PCR products are detected
using the ABI 7700 Sequence Detection System. In some embodiments,
an agent that inhibits cell lysis has been added to the sample to
inhibit the lysis of cells, if present, where the agent is membrane
stabilizer, cross-linker, or cell lysis inhibitor. In some
embodiments, the agent is a cell lysis inhibitor. In some of these
embodiments, the cell lysis inhibitor is formalin at a percentage
selected from the group consisting of: 0.0001-0.03%, 0.03-0.05%,
0.05-0.08%, 0.08-0.1%, 0.1-0.3%, 0.3-0.5%, 0.5-0.7%, 0.7-0.9%,
0.9-1.2%, 1.2-1.5%, 1.5-2%, or 2-3%. In some embodiments, the
concentration of formalin in the sample is 0.1%.
[0062] In another aspect, the invention provides methods for
determining the sequence of a locus of interest in a sample
containing fetal DNA.
[0063] In some embodiments, the method for determining the sequence
includes (a) amplifying a locus of interest on a template DNA using
a first and second primers, where the second primer contains a
recognition site for a restriction enzyme such that digestion with
the restriction enzyme generates a 5' overhang containing the locus
of interest; (b) digesting the amplified DNA with the restriction
enzyme that recognizes the recognition site on the second primer;
(c) incorporating a nucleotide into the digested DNA of (b) by
using the 5' overhang containing the locus of interest as a
template; and (d) determining the sequence of the locus of interest
by determining the sequence of the DNA of (c).
[0064] In other embodiments, the method for determining the
sequence includes (a) amplifying alleles of a locus of interest on
a template DNA using a first and second primers, where the second
primer contains a recognition site for a restriction enzyme such
that digestion with the restriction enzyme generates a 5' overhang
containing the locus of interest; (b) digesting the amplified DNA
with the restriction enzyme that recognizes the recognition site on
the second primer; (c) incorporating nucleotides into the digested
DNA of (b), where a nucleotide that terminates elongation, and is
complementary to the locus of interest of an allele, is
incorporated into the 5' overhang of said allele, and a nucleotide
complementary to the locus of interest of a different allele is
incorporated into the 5' overhang of said different allele, and
said terminating nucleotide, which is complementary to a nucleotide
in the 5' overhang of said different allele, is incorporated into
the 5' overhang of said different allele; and (d) determining the
sequence of the alleles of a locus of interest by determining the
sequence of the DNA of (c).
[0065] In embodiments, the sample is cell, tissue, blood, serum,
plasma, saliva, urine, tears, vaginal secretion, sweat, umbilical
cord blood, chorionic villi, amniotic fluid, embryonic tissue,
embryo, a two-celled embryo, a four-celled embryo, an eight-celled
embryo, a 16-celled embryo, a 32-celled embryo, a 64-celled embryo,
a 128-celled embryo, a 256-celled embryo, a 512-celled embryo, a
1024-celled embryo, lymph fluid, cerebrospinal fluid, mucosa
secretion, peritoneal fluid, ascitic fluid, fecal matter, or body
exudates.
[0066] In another aspect, the invention provides methods for
preparing a sample for analysis that include isolating free nucleic
acid from a sample that contains nucleic acid, where an agent that
inhibits cell lysis has been added to the sample to inhibit lysis
of cells, if cells are present, where the agent is membrane
stabilizer, cross-linker, or cell lysis inhibitor. In this aspect,
the portion of the sample that is to be analyzed is the free
nucleic acid, not the cellular portion. In an embodiment, the
present invention provides a method for isolating nucleic acid said
method comprising (a) obtaining a sample containing nucleic acid;
(b) adding a cell lysis inhibitor, cell membrane stabilizer, or
cross-linker to the sample of (a); and (c) isolating nucleic acid.
In an embodiment, the method is used for isolating free nucleic
acid. In an embodiment, the method is used for isolating free fetal
nucleic acid. In another embodiment, the present invention provides
a method for isolating free fetal nucleic acid said method
comprising (a) obtaining a sample containing nucleic acid; (b)
adding a cell lysis inhibitor, cell membrane stabilizer, or
cross-linker to the sample of (a); (c) isolating the plasma from
the blood sample, wherein the plasma is isolated by centrifuging
the blood sample; and (d) removing the supernatant, which contains
the plasma, using procedures to minimize disruption of the
"buffy-coat."
[0067] In some embodiments, the agent is cell lysis inhibitor, and
in some of these embodiments, the cell lysis inhibitor is
glutaraldehyde, derivatives of glutaraldehyde, formaldehyde,
formalin, and derivatives of formaldehyde, crosslinkers, primary
amine reactive crosslinkers, sulfhydryl reactive crosslinkers,
sulfhydryl addition or disulfide reduction, carbohydrate reactive
crosslinkers, carboxyl reactive crosslinkers, photoreactive
crosslinkers, cleavable crosslinkers, AEDP, APG, BASED,
BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3, BSOCOES,
DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP, DTME, DTSSP,
EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS or compounds listed
in Table XXIII.
[0068] In some embodiments the cell lysis inhibitor is formalin. In
some of these embodiments, the final concentration of formalin in
the sample is 0.0001-0.03%, 0.03-0.05%, 0.05-0.08%, 0.08-0.1%,
0.1-0.3%, 0.3-0.5%, 0.5-0.7%, 0.7-0.9%, 0.9-1.2%, 1.2-1.5%, 1.5-2%,
or 2-3%. In one embodiment, the final concentration of formalin in
the sample is 0.1%.
[0069] An agent that stabilizes cell membranes may be added to the
sample including but not limited to aldehydes, urea formaldehyde,
phenol formaldehyde, DMAE (dimethylaminoethanol), cholesterol,
cholesterol derivatives, high concentrations of magnesium, vitamin
E, and vitamin E derivatives, calcium, calcium gluconate, taurine,
niacin, hydroxylamine derivatives, bimoclomol, sucrose,
astaxanthin, glucose, amitriptyline, isomer A hopane tetral
phenylacetate, isomer B hopane tetral phenylacetate, citicoline,
inositol, vitamin B, vitamin B complex, cholesterol hemisuccinate,
sorbitol, calcium, coenzyme Q, ubiquinone, vitamin K, vitamin K
complex, menaquinone, zonegran, zinc, ginkgo biloba extract,
diphenylhydantoin, perftoran, polyvinylpyrrolidone,
phosphatidylserine, tegretol, PABA, disodium cromglycate,
nedocromil sodium, phenyloin, zinc citrate, mexitil, dilantin,
sodium hyaluronate, or polaxamer 188.
[0070] In another embodiment, an agent that prevents DNA
destruction is added to the sample including but not limited to
DNase inhibitors, zinc chloride, ethylenediaminetetraacetic acid,
guanidine-HCl, guanidine isothiocyanate, N-lauroylsarcosine, and
Na-dodecylsulphate.
[0071] In some embodiments, the sample is obtained from human,
non-human, mammal, reptile, cattle, cat, dog, goat, swine, pig,
monkey, ape, gorilla, bull, cow, bear, horse, sheep, poultry,
mouse, rat, fish, dolphin, whale, or shark. In some of these
embodiments, the sample is obtained from a human source.
[0072] In some embodiments, the sample containing nucleic acid is
obtained from any nucleic acid containing source including but not
limited to a cell, fetal cell, tissue, blood, serum, plasma,
saliva, urine, tear, vaginal secretion, breast fluid, breast milk,
sweat, umbilical cord blood, chorionic villi, amniotic fluid,
embryonic tissue, embryo, a two-celled embryo, a four-celled
embryo, an eight-celled embryo, a 16-celled embryo, a 32-celled
embryo, a 64-celled embryo, a 128-celled embryo, a 256-celled
embryo, a 512-celled embryo, a 1024-celled embryo lymph fluid,
cerebrospinal fluid, mucosa secretion, peritoneal fluid, ascitic
fluid, fecal matter, or body exudates. In some of these
embodiments, the sample is blood.
[0073] In embodiments the sample is from a pregnant female. In an
embodiment, the sample is obtained from a pregnant human female. In
an embodiment, the sample is blood obtained from a pregnant female
and, e.g., the nucleic acid is isolated from plasma obtained from
blood of a pregnant female; the plasma is generated using
procedures designed to minimize the amount of maternal cell lysis.
In some of these embodiments, the blood is obtained from a human
pregnant female when the fetus is at a gestational age of 0-4, 4-8,
8-12, 12-16, 16-20, 20-24, 24-28, 28-32, 32-36, 36-40, 40-44,
44-48, 48-52, or more than 52 weeks. In some of these embodiments,
the sample is obtained from plasma from the blood.
[0074] In some embodiments, the isolation of nucleic acid includes
a centrifugation step; e.g., in some embodiments free nucleic acid
is isolated from plasma obtained from blood, for example from a
pregnant female. In some embodiments, the centrifugation step is
performed with the centrifuge braking power set to zero (the
centrifuge comes to a stop by natural deceleration). In some
embodiments, the centrifugation step is performed at a speed of
0-50 rpm, 50-100 rpm, 100-200 rpm, 200-300 rpm, 300-400 rpm,
400-500 rpm, 500-600 rpm, 600-700 rpm, 700-800 rpm, 800-900 rpm,
900-1000 rpm, 1000-2000 rpm, 2000-3000 rpm, 3000-4000 rpm,
4000-5000 rpm, 5000-6000 rpm, 6000-7000 rpm, 7000-8000 rpm, or
greater than 8000 rpm. In one embodiment, the blood, e.g., from the
pregnant female, is centrifuged at a speed less than 4000 rpm. In
another embodiment, the acceleration power of the centrifuge is not
used.
[0075] In another aspect, the invention provides a method for
detecting a chromosomal abnormality by (a) determining the sequence
of alleles of a locus of interest from template DNA, and (b)
quantitating the relative amount of the alleles at a heterozygous
locus of interest that was identified from the locus of interest of
(a), wherein said relative amount is expressed as a ratio, and
wherein said ratio indicates the presence or absence of a
chromosomal abnormality.
[0076] In yet another aspect, the invention provides
compositions,
[0077] In one embodiment, the invention provides a composition
containing fetal DNA and maternal DNA, where the percentage of free
fetal DNA in the total free DNA of the composition is about 15-16%
fetal DNA, about 16-17% fetal DNA, about 17-18% fetal DNA, about
18-19% fetal DNA, about 19-20% fetal DNA, about 20-21% fetal DNA,
about 21-22% fetal DNA, about 22-23% fetal DNA, about 23-24% fetal
DNA, about 24-25% fetal DNA, about 25-35% fetal DNA, about 35-45%
fetal DNA, about 45-55% fetal DNA, about 55-65% fetal DNA, about
65-75% fetal DNA, about 75-85% fetal DNA, about 85-90% fetal DNA,
about 90-91% fetal DNA, about 91-92% fetal DNA, about 92-93% fetal
DNA, about 93-94% fetal DNA, about 94-95% fetal DNA, about 95-96%
fetal DNA, about 96-97% fetal DNA, about 97-98% fetal DNA, about
98-99% fetal DNA, or about 99-99.7% fetal DNA.
[0078] In another embodiment, the invention provides a composition
containing fetal DNA and maternal DNA, where the percentage of free
fetal DNA in the total free DNA of the composition is about 15-16%
fetal DNA, about 16-17% fetal DNA, about 17-18% fetal DNA, about
18-19% fetal DNA, about 19-20% fetal DNA, about 20-21% fetal DNA,
about 21-22% fetal DNA, about 22-23% fetal DNA, about 23-24% fetal
DNA, about 24-25% fetal DNA, about 25-35% fetal DNA, about 35-45%
fetal DNA, about 45-55% fetal DNA, about 55-65% fetal DNA, about
65-75% fetal DNA, about 75-85% fetal DNA, about 85-90% fetal DNA,
about 90-91% fetal DNA, about 91-92% fetal DNA, about 92-93% fetal
DNA, about 93-94% fetal DNA, or about 94-95% fetal DNA.
[0079] In yet another aspect, the invention provides a prenatal
diagnostic method including analyzing a composition comprising
fetal DNA and maternal DNA, where the percentage of free fetal DNA
in the total free DNA of the composition is about 15-16% fetal DNA,
about 16-17% fetal DNA, about 17-18% fetal DNA, about 18-19% fetal
DNA, about 19-20% fetal DNA, about 20-21% fetal DNA, about 21-22%
fetal DNA, about 22-23% fetal DNA, about 23-24% fetal DNA, about
24-25% fetal DNA, about 25-35% fetal DNA, about 35-45% fetal DNA,
about 45-55% fetal DNA, about 55-65% fetal DNA, about 65-75% fetal
DNA, about 75-85% fetal DNA, about 85-90% fetal DNA, about 90-91%
fetal DNA, about 91-92% fetal DNA, about 92-93% fetal DNA, about
93-94% fetal DNA, or about 94-95% fetal DNA.
[0080] In still yet another aspect, the invention provides a kit
for use in any of the methods of the invention, where the kit
contains a set of primers used in the method, where the second
primer contains a sequence that generates a recognition site for a
restriction enzyme such that digestion with the restriction enzyme
generates a 5' overhang containing the locus of interest, and a set
of instructions.
BRIEF DESCRIPTION OF THE FIGURES
[0081] FIG. 1A. A schematic diagram depicting a double stranded DNA
molecule. A pair of primers, depicted as bent arrows, flank the
locus of interest, depicted as a triangle symbol at base N14. The
locus of interest can be a single nucleotide polymorphism, point
mutation, insertion, deletion, translocation, etc. Each primer
contains a restriction enzyme recognition site about 10 by from the
5' terminus depicted as region "a" in the first primer and as
region "d" in the second primer, Restriction recognition site "a"
can be for any type of restriction enzyme but recognition site "d"
is for a restriction enzyme, which cuts "n" nucleotides away from
its recognition site and leaves a 5' overhang and a recessed 3'
end. Examples of such enzymes include but are not limited to BceAI
and BsmF I. The 5' overhang serves as a template for incorporation
of a nucleotide into the 3' recessed end.
[0082] The first primer is shown modified with biotin at the 5' end
to aid in purification. The sequence of the 3' end of the primers
is such that the primers anneal at a desired distance upstream and
downstream of the locus of interest. The second primer anneals
close to the locus of interest; the annealing site, which is
depicted as region "c," is designed such that the 3' end of the
second primer anneals one base away from the locus of interest. The
second primer can anneal any distance from the locus of interest
provided that digestion with the restriction enzyme, which
recognizes the region "d" on this primer, generates a 5' overhang
that contains the locus of interest. The first primer annealing
site, which is depicted as region "b," is about 20 bases.
[0083] FIG. 1B. A schematic diagram depicting the annealing and
extension steps of the first cycle of amplification by PCR. The
first cycle of amplification is performed at about the melting
temperature of the 3' region, which anneals to the template DNA, of
the second primer, depicted as region "c," and is 13 base pairs in
this example. At this temperature, both the first and second
primers anneal to their respective complementary strands and begin
extension, depicted by dotted lines. In this first cycle, the
second primer extends and copies the region b where the first
primer can anneal in the next cycle.
[0084] FIG. 1C. A schematic diagram depicting the annealing and
extension steps following denaturation in the second cycle of
amplification of PCR. The second cycle of amplification is
performed at a higher annealing temperature (TM2), which is about
the melting temperature of the 20 by of the 3' region of the first
primer that anneals to the template DNA, depicted as region "b."
Therefore at TM2, the first primer, which contains region b' which
is complementary to region b, can bind to the DNA that was copied
in the first cycle of the reaction. However, at TM2 the second
primer cannot anneal to the original template DNA or to DNA that
was copied in the first cycle of the reaction because the annealing
temperature is too high. The second primer can anneal to 13 bases
in the original template DNA but TM2 is calculated at about the
melting temperature of 20 bases.
[0085] FIG. 1D. A schematic diagram depicting the annealing and
extension reactions after denaturation during the third cycle of
amplification. In this cycle, the annealing temperature, TM3, is
about the melting temperature of the entire second primer,
including regions "c" and "d." The length of regions "c"+"d" is
about 27-33 by long, and thus TM3 is significantly higher than TM1
and TM2. At this higher TM the second primer, which contain regions
c' and d', anneals to the copied DNA generated in cycle 2.
[0086] FIG. 1E. A schematic diagram depicting the annealing and
extension reactions for the remaining cycles of amplification. The
annealing temperature for the remaining cycles is TM3, which is
about the melting temperature of the entire second primer. At TM3,
the second primer binds to templates that contain regions c' and d'
and the first primer binds to templates that contain regions a' and
b. By raising the annealing temperature successively in each cycle
for the first three cycles, from TM1, TM2, and TM3, nonspecific
amplification is significantly reduced.
[0087] FIG. 1F. A schematic diagram depicting the amplified locus
of interest bound to a solid matrix.
[0088] FIG. 1G. A schematic diagram depicting the bound, amplified.
DNA after digestion with restriction enzyme "d." The "downstream"
end is released into the supernatant, and can be removed by washing
with any suitable buffer. The upstream end containing the locus of
interest remains bound to the solid matrix.
[0089] FIG. 1H. A schematic diagram depicting the bound amplified
DNA, after "filling in" with a labeled ddNTP. A DNA polymerase is
used to "fill in" the base (N'14) that is complementary to the
locus of interest (N14). In this example, only ddNTPs are present
in this reaction, such that only the locus of interest or SNP of
interest is filled in.
[0090] FIG. 1I. A schematic diagram depicting the labeled, bound
DNA after digestion with restriction enzyme "a." The labeled DNA is
released into the supernatant, which can be collected to identify
the base that was incorporated.
[0091] FIG. 2. A schematic diagram depicting double stranded DNA
templates with n number of loci of interest and n number of primer
pairs, x.sub.1, y.sub.1 to x.sub.n, y.sub.n, specifically annealed
such that a primer flanks each locus of interest. The first primers
are biotinylated at the 5' end, depicted by *, and contain a
restriction enzyme recognition site, "a", which can be any type of
restriction enzyme. The second primers contain a restriction enzyme
recognition site, "d," where "d" is a recognition site for a
restriction enzyme that cuts "n" nucleotides away from its
recognition site, and generates a 5' overhang containing the locus
of interest and a recessed 3' end. The second primers anneal
adjacent to the respective loci of interest. The exact position of
the restriction enzyme site "d" in the second primers is designed
such that digesting the PCR product of each locus of interest with
restriction enzyme "d" generates a 5' overhang containing the locus
of interest and a 3' recessed end. The annealing sites of the first
primers are about 20 bases long and are selected such that each
successive first primer is further away from its respective second
primer. For example, if at locus 1 the 3' ends of the first and
second primers are Z base pairs apart, then at locus 2, the 3' ends
of the first and second primers are Z+K base pairs apart, where
K=1, 2, 3 or more than three bases. Primers for locus N are
Z.sub.N-1+K base pairs apart. The purpose of making each successive
first primer further apart from their respective second primers is
such that the "filled in" restriction fragments (generated after
amplification, purification, digestion and labeling as described in
FIGS. 1B-1I) differ in size and can be resolved, for example by
electrophoresis, to allow detection of each individual locus of
interest.
[0092] FIG. 3A-3C: PCR amplification of SNPs using multiple
annealing temperatures. A sample containing genomic DNA templates
from thirty-six human volunteers was analyzed for the following
four SNPs: SNP HC21S00340 (lane 1), identification number as
assigned in the Human Chromosome 21 cSNP Database, located on
chromosome 21; SNP TSC 0095512 (lane 2), located on chromosome 1,
SNP TSC 0214366 (lane 3), located on chromosome 1; and SNP TSC
0087315 (lane 4), located on chromosome 1. Each SNP was amplified
by PCR using three different annealing temperature protocols,
herein referred to as the low stringency annealing temperature;
medium stringency annealing temperature; and high stringency
annealing temperature. Regardless of the annealing temperature
protocol, each SNP was amplified for 40 cycles of PCR. The
denaturation step for each PCR reaction was performed for 30
seconds at 95.degree. C. 3A. Photograph of a gel demonstrating PCR
amplification of the 4 different SNPs using the low stringency
annealing temperature protocol. 3B. Photograph of a gel
demonstrating PCR amplification of the 4 different SNPs using
medium stringency annealing temperature protocol. 3C. Photograph of
a gel demonstrating PCR amplification of the 4 different SNPs using
the high stringency annealing temperature protocol.
[0093] FIG. 4A. (From top to bottom: SEQ ID NOS: 17, 667, 668, 18.)
A depiction of the DNA sequence of SNP HC21S00027, as assigned by
the Human Chromosome 21 cSNP database, located on chromosome 21. A
first primer and a second primer are indicated above and below,
respectively, the sequence of HC21S00027. The first primer is
biotinylated and contains the restriction enzyme recognition site
for EcoRI. The second primer contains the restriction enzyme
recognition site for BsmF I and contains 13 bases that anneal to
the DNA sequence. The SNP is indicated by R (A/G) and r (T/C)
(complementary to R).
[0094] FIG. 4B. (From top to bottom: SEQ ID NOS: 17, 667, 668, 19.)
A depiction of the DNA sequence of SNP HC21S00027, as assigned by
the Human Chromosome 21 cSNP database, located on chromosome 21. A
first primer and a second primer are indicated above and below,
respectively, the sequence of HC21S00027. The first primer is
biotinylated and contains the restriction enzyme recognition site
for EcoRI. The second primer contains the restriction enzyme
recognition site for BceA I and has 13 bases that anneal to the DNA
sequence. The SNP is indicated by R (A/G) and r (T/C)
(complementary to R).
[0095] FIG. 4C. (From top to bottom: SEQ ID NOS: 11, 669, 670, 20.)
A depiction of the DNA sequence of SNP TSC0095512 from chromosome
1. The first primer and the second primer are indicated above and
below, respectively, the sequence of TSC0095512. The first primer
is biotinylated and contains the restriction enzyme recognition
site for EcoRI. The second primer contains the restriction enzyme
recognition site for BsmF I and has 13 bases that anneal to the DNA
sequence. The SNP is indicated by S (G/C) and s (C/G)
(complementary to S).
[0096] FIG. 4D. (From top to bottom: SEQ ID NOS: 11, 669, 670, 12.)
A depiction of the DNA sequence of SNP TSC0095512 from chromosome
1. The first primer and the second primer are indicated above and
below, respectively, the sequence of TSC0095512. The first primer
is biotinylated and contains the restriction enzyme recognition
site for EcoRI. The second primer contains the restriction enzyme
recognition site for BceA I and has 13 bases that anneal to the DNA
sequence. The SNP is indicated by S (G/C) and s (C/G)
(complementary to S).
[0097] FIGS. 5A-5D. (FIG. 5A: SEQ ID NOS: 671 (top) and 672
(bottom); FIG. 5B: SEQ ID NOS: 673 (top) and 674 (bottom); FIG. 5C:
SEQ ID NOS: 675 (top) and 676 (bottom); FIG. 5D: SEQ ID NOS: 677
(top) and 678 (bottom)). A schematic diagram depicting the
nucleotide sequences of SNP HC21S00027 (FIGS. 5A and 5B) and SNP
TSC0095512 (FIGS. 5C and 5D) after amplification with the primers
described in FIGS. 4A-4D. Restriction sites in the primer sequence
are indicated in bold.
[0098] FIGS. 6A-6D. A schematic diagram depicting the nucleotide
sequences of each amplified SNP after digestion with the
appropriate Type IIS restriction enzyme. FIGS. 6A (SEQ ID NOS: 679
(upper left), 680 (upper right), 681 (lower left) and 682 (lower
right)) and 6B (SEQ ID NOS: 679 (upper left), 683 (upper right),
684 (lower left) and 685 (lower right)) depict fragments of SNP
HC21S00027 digested with the Type IIS restriction enzymes BsmF I
and BceA I, respectively. FIGS. 6C (SEQ ID NOS: 686 (upper left),
687 (upper right), 688 (lower left) and 689 (lower right)) and 6D
(SEQ ID NOS: 686 (upper left), 690 (upper right), 691 (lower left)
and 692 (lower right)) depict fragments of SNP TSC0095512 digested
with the Type IIS restriction enzymes BsmF I and BceA I,
respectively.
[0099] FIGS. 7A-7D. A schematic diagram depicting the incorporation
of a fluorescently labeled nucleotide using the 5' overhang of the
digested SNP site as a template to "fill in" the 3' recessed end.
FIGS. 7A (SEQ ID NOS: 693 (top) and 694 (bottom)) and 7B (SEQ ID
NOS: 693 (top) and 695 (bottom)) depict the digested SNP HC21S00027
locus with an incorporated labeled ddNTP (*R.sup.-dd=fluorescent
dideoxy nucleotide). FIGS. 7C (SEQ ID NOS: 696 (top) and 697
(bottom)) and 7D (SEQ ID NOS: 696 (top) and 698 (bottom)) depict
the digested SNP TSC0095512 locus with an incorporated labeled
ddNTP (*S.sup.-dd=fluorescent dideoxy nucleotide). The use of
ddNTPs ensures that the 3' recessed end is extended by one
nucleotide, which is complementary to the nucleotide of interest or
SNP site present in the 5' overhang.
[0100] FIG. 7E. (From top to bottom: SEQ ID NOS: 693, 694, 699,
694, 700, 694, 701, 694.) A schematic diagram depicting the
incorporation of dNTPs and a ddNTP into the 5' overhang containing
the SNP site. SNP HC21500007 was digested with BsmF I, which
generates a four base 5' overhang. The use of a mixture of dNTPs
and ddNTPs allows the 3' recessed end to be extended one nucleotide
(a ddNTP is incorporated first); two nucleotides (a dNTP is
incorporated followed by a ddNTP); three nucleotides (two dNTPs are
incorporated, followed by a ddNTP); or four nucleotides (three
dNTPs are incorporated, followed by a ddNTP). All four products can
be separated by size, and the incorporated nucleotide detected
(*R.sup.-dd=fluorescent dideoxy nucleotide). Detection of the first
nucleotide, which corresponds to the SNP or locus site, and the
next three nucleotides provides an additional level of quality
assurance. The SNP is indicated by R (A/G) and r (T/C)
(complementary to R).
[0101] FIGS. 8A-8D. Release of the "filled in" SNP from the solid
support matrix, i.e. streptavidin coated well. SNP HC21S00027 is
shown in FIGS. 8A (SEQ ID NOS: 702 (upper left), 703 (upper right),
704 (lower left) and 705 (lower right)) and 8B (SEQ ID NOS: 702
(upper left), 703 (upper right), 704 (lower left) and 706 (lower
right)), while SNP TSC0095512 is shown in FIGS. 8C (SEQ ID NOS: 707
(upper left), 708 (upper right), 709 (lower left) and 710 (lower
right)) and 8D (SEQ ID NOS: 707 (upper left), 708 (upper right),
709 (lower left) and 711 (lower right)). The "filled in" SNP is
free in solution, and can be detected.
[0102] FIG. 9A. Sequence analysis of SNP HC21S00027 digested with
BceAI. Four "fill in" reactions are shown; each reaction contained
one fluorescently labeled nucleotide, ddGTP, ddATP, ddTTP, or
ddCTP, and unlabeled ddNTPs. The 5' overhang generated by digestion
with BceA I and the expected nucleotides at this SNP site are
indicated.
[0103] FIG. 9B. Sequence analysis of SNP TSC0095512. SNP TSC0095512
was amplified with a second primer that contained the recognition
site for BceA I, and in a separate reaction, with a second primer
that contained the recognition site for BsmF I. Four fill in
reactions are shown for each PCR product; each reaction contained
one fluorescently labeled nucleotide, ddGTP, ddATP, ddTTP, or
ddCTP, and unlabeled ddNTPs. The 5' overhang generated by digestion
with BceA I and with BsmF I and the expected nucleotides are
indicated.
[0104] FIG. 9C. Sequence analysis of SNP TSC0264580 after
amplification with a second primer that contained the recognition
site for BsmF I. Four fill in reactions are shown; each reaction
contained one fluorescently labeled nucleotide, which was ddGTP,
ddATP, ddTTP, or ddCTP and unlabeled ddNTPs. Two different 5'
overhangs are depicted: one represents the DNA molecules that were
cut 11 nucleotides away on the sense strand and 15 nucleotides away
on the antisense strand and the other represents the DNA molecules
that were cut 10 nucleotides away on the sense strand and 14
nucleotides away on the antisense strand. The expected nucleotides
also are indicated.
[0105] FIG. 9D. Sequence analysis of SNP HC21 S00027 amplified with
a second primer that contained the recognition site for BsmF I. A
mixture of labeled ddNTPs and unlabeled dNTPs was used to fill in
the 5' overhang generated by digestion with BsmF I. Two different
5' overhangs are depicted: one represents the DNA molecules that
were cut 11 nucleotides away on the sense strand and 15 nucleotides
away on the antisense strand and the other represents the DNA
molecules that were cut 10 nucleotides away on the sense strand and
14 nucleotides away on the antisense strand. The nucleotide
upstream from the SNP, the nucleotide at the SNP site (the sample
contained DNA templates from 36 individuals; both nucleotides would
be expected to be represented in the sample), and the three
nucleotides downstream of the SNP are indicated.
[0106] FIG. 10. Sequence analysis of multiple SNPs. SNPs
HC21S00131, and HC21S00027, which are located on chromosome 21, and
SNPs TSC0087315, SNP TSC0214366, SNP TSC0413944, and SNP
TSC0095512, which are on chromosome 1, were amplified in separate
PCR reactions with second primers that contained a recognition site
for BsmF I. The primers were designed so that each amplified locus
of interest was of a different size. After amplification, the
reactions were pooled into a single sample, and all subsequent
steps of the method performed (as described for FIGS. 1F-1I) on
that sample. Each SNP and the nucleotide found at each SNP are
indicated.
[0107] FIG. 11A-11B. Quantification of the percentage of fetal DNA
in maternal blood. Blood was obtained from a pregnant human female
with informed consent. DNA was isolated and serial dilutions were
made to determine the percentage of fetal DNA present in the
sample. The SRY gene, which is located on chromosome Y, was used to
detect fetal DNA. The cystic fibrosis gene, which is located on
chromosome 7, was used to detect both maternal and fetal DNA. 11 A.
Amplification of the SRY gene and the cystic fibrosis gene using a
DNA template isolated from a blood sample that was treated with
EDTA. 11B. Amplification of the SRY gene and the cystic fibrosis
gene using a DNA template that was isolated from a blood sample
that was treated with formalin and EDTA.
[0108] FIG. 12. Genetic analysis of an individual previously
genotyped with Trisomy 21 (Down's Syndrome). Blood was collected,
with informed consent, from an individual who had previously been
genotyped with trisomy 21. DNA was isolated and two SNPs on
chromosome 21 and two SNPs on chromosome 13 were genotyped. As
shown in the photograph of the gel, the SNPs at chromosome 21 show
disproportionate ratios of the two nucleotides. Visual inspection
of the gel demonstrates that one nucleotide of the two nucleotides
at the SNP sites analyzed for chromosome 21 is of greater
intensity, suggesting it is not present in a 50:50 ratio. However,
visual inspection of the gel suggests that the nucleotides at the
heterozygous SNP sites analyzed on chromosome 13 are present in the
expected 50:50 ratio.
[0109] FIG. 13. Sequence determination of both alleles of SNPs
TSC0837969, TSC0034767, TSC1130902, TSC0597888, TSC0195492,
TSC0607185 using one fluorescently labeled nucleotide. Labeled
ddGTP was used in the presence of unlabeled dATP, dCTP, dTTP to
fill-in the overhang generated by digestion with BsmF I. The
nucleotide preceding the variable site on the strand that was
filled-in was not guanine, and the nucleotide after the variable
site on the strand that was filled in was not guanine. The
nucleotide two bases after the variable site on the strand that was
filled-in was guanine. Alleles that contain guanine at variable
site are filled in with labeled ddGTP. Alleles that do not contain
guanine are filled in with unlabeled dATP, dCTP, or dTTP, and the
polymerase continues to incorporate nucleotides until labeled ddGTP
is filled in at position 3 complementary to the overhang.
[0110] FIG. 14. Identification of SNPs with alleles that are
variable within the population. The sequences of both alleles of
seven SNPs located on chromosome 13 were determined using a
template DNA comprised of DNA obtained from two hundred and forty
five individuals. Labeled ddGTP was used in the presence of
unlabeled dATP, dCTP, dTTP to fill-in the overhang generated by
digestion with BsmF I. The nucleotide preceding the variable site
on the strand that was filled-in was not guanine, and the
nucleotide after the variable site on the strand that was filled in
was not guanine. The nucleotide two bases after the variable site
on the strand that was filled-in was guanine. Alleles that contain
guanine at variable site are filled in with labeled ddGTP. Alleles
that do not contain guanine are filled in with unlabeled dATP,
dCTP, or dTTP, and the polymerase continues to incorporate
nucleotides until labeled ddGTP is filled in at position 3
complementary to the overhang.
[0111] FIG. 15. Determination of the ratio for one allele to the
other allele at heterozygous SNPs. The observed nucleotides for SNP
TSC0607185 are cytosine (referred to as allele 1) and thymidine
(referred to as allele 2) on the sense strand. The ratio of allele
2 to allele 1 was calculated using template DNA isolated from five
individuals. The ratio of allele 2 to allele 1 (allele 2/allele 1)
was consistently 1:1.
[0112] The observed nucleotides for SNP TSC1130902 are guanine
(referred to as allele 1) and adenine (referred to as allele 2) on
the sense strand. The ratio of allele 2 to allele 1 was calculated
using template DNA isolated from five individuals. The ratio of
allele 2 to allele 1 (allele 2/allele 1) was consistently
75:25.
[0113] FIG. 16. The percentage of allele 2 to allele 1 at SNP
TSC0108992 remains linear when calculated on template DNA
containing an extra copy of chromosome 21. SNP TSC0108992 was
amplified using template DNA from four individuals, and two
separate fill-in reactions (labeled as A and B) were performed for
each PCR reaction (labeled 1 through 4). The calculated percentage
of allele 2 to allele 1 on template DNA from normal individuals was
0.47. The deviation from the theoretically predicted percentage of
0.50 remained linear on template DNA isolated from an individual
with Down's syndrome.
[0114] FIG. 17A. Analysis of a SNP located on chromosome 21 from
template DNA isolated from an individual with a normal genetic
karyotype. SNP TSC0108992 was amplified using the methods described
herein, and after digestion with the type IIS restriction enzyme
BsmF I, the 5' overhang was filled in using labeled ddTTP, and
unlabeled dATP, dCTP, and dGTP. Three separate PCR reactions were
performed, and each PCR reaction was split into two samples. The
percentage of allele 2 at the SNP site (allele 2/(allele 2+allele
1)) was calculated, which resulted in mean of 0.50.
[0115] FIG. 17B. Analysis of a SNP located on chromosome 21 from
template DNA isolated from an individual with a trisomy 21 genetic
karyotype. SNP TSC0108992 was amplified using the methods described
herein, and after digestion with the type IIS restriction enzyme
BsmF I, the 5' overhang was filled in using labeled ddTTP, and
unlabeled dATP, dCTP, and dGTP. Three separate PCR reactions were
performed, and each PCR reaction was split into two samples. The
percentage of allele 2 at the SNP site (allele 2/(allele 2+allele
1)) was calculated, which resulted in mean of 0.30.
[0116] FIG. 17C. Analysis of a SNP located on chromosome 21 from a
mixture comprised of template DNA from an individual with Trisomy
21, and template DNA from an individual with a normal genetic
karyotype in a ratio of 3:1 (Trisomy 21: Normal). SNP TSC0108992
was amplified from the mixture of template DNA using the methods
described herein, and after digestion with the type IIS restriction
enzyme BsmF I, the 5' overhang was filled in using labeled ddTTP,
and unlabeled dATP, dCTP, and dGTP. Three separate PCR reactions
were performed, and each PCR reaction was split into two samples.
The percentage of allele 2 at the SNP site (allele 2/(allele
2+allele 1)) was calculated, which resulted in mean of 0.319.
[0117] FIG. 17D. Analysis of a SNP located on chromosome 21 from a
mixture comprised of template DNA from an individual with Trisomy
21, and template DNA from an individual with a normal genetic
karyotype in a ratio of 1:1 (Trisomy 21: Normal). SNP TSC0108992
was amplified from the mixture of template DNA using the methods
described herein, and after digestion with the type IIS restriction
enzyme BsmF I, the 5' overhang was filled in using labeled ddTTP,
and unlabeled dATP, dCTP, and dGTP. Three separate PCR reactions
were performed, and each PCR reaction was split into two samples.
The percentage of allele 2 at the SNP site (allele 2/(allele
2+allele 1)) was calculated, which resulted in mean of 0.352.
[0118] FIG. 17E. Analysis of a SNP located on chromosome 21 from a
mixture comprised of template DNA from an individual with Trisomy
21, and template DNA from an individual with a normal genetic
karyotype in a ratio of 1:2.3 (Trisomy 21: Normal). SNP TSC0108992
was amplified from the mixture of template DNA using the methods
described herein, and after digestion with the type IIS restriction
enzyme BsmF I, the 5' overhang was filled in using labeled ddTTP,
and unlabeled dATP, dCTP, and dGTP. Three separate PCR reactions
were performed, and each PCR reaction was split into two samples.
The percentage of allele 2 at the SNP site (allele 2/(allele
2+allele 1)) was calculated, which resulted in mean of 0.382.
[0119] FIG. 17F. Analysis of a SNP located on chromosome 21 from a
mixture comprised of template DNA from an individual with Trisomy
21, and template DNA from an individual with a normal genetic
karyotype in a ratio of 1:4 (Trisomy 21: Normal). SNP TSC0108992
was amplified from the mixture of template DNA using the methods
described herein, and after digestion with the type IIS restriction
enzyme BsmF I, the 5' overhang was filled in using labeled ddTTP,
and unlabeled dATP, dCTP, and dGTP. Three separate PCR reactions
were performed, and each PCR reaction was split into two samples.
The percentage of allele 2 at the SNP site (allele 2/(allele
2+allele 1)) was calculated, which resulted in mean of 0.397.
[0120] FIG. 18A. Agarose gel analysis of nine (9) SNPs amplified
from template DNA. Each of the nine SNPs were amplified from
genomic DNA using the methods described herein. Lane 1 corresponds
to SNP TSC0397235, lane 2 corresponds to TSC0470003, lane 3
corresponds to TSC1649726, lane 4 corresponds to TSC1261039, lane 5
corresponds to TSC0310507, lane 6 corresponds to TSC1650432, lane 7
corresponds to TSC1335008, lane 8 corresponds to TSC0128307, and
lane 9 corresponds to TSC0259757.
[0121] FIG. 18B. The original template DNA was amplified using 12
base primers that annealed to various regions on chromosome 13. One
hundred different primer sets were used to amplify regions
throughout chromosome 13. For each of the nine SNPs, a primer that
annealed approximately 130 bases from the locus of interest and 130
bases downstream of the locus of interest were used. This
amplification reaction, which contained a total of 100 different
primer sets, was used to amplify the regions containing the loci of
interest. The resulting PCR product was used in a subsequent PCR
reaction, wherein each of the nine SNPs were individually amplified
using a first primer and a second primer, wherein the second primer
contained the binding site for the type IIs restriction enzyme BsmF
I. SNPs were loaded in the same order as FIG. 18A.
[0122] FIG. 19A. Quantification of the percentage of allele 2 to
allele 1 for SNP TSC047003 on original template DNA (IA) and
multiplexed template DNA (M1-M3), wherein the DNA was first
amplified using 12 base primers that annealed 150 bases upstream
and downstream of the loci of interest. Then, three separate PCR
reactions were performed on the multiplexed template DNA, using a
first and second primer.
[0123] FIG. 19B. Quantification of the percentage of allele 2 to
allele 1 for SNP TSC1261039 on original template DNA (IA) and
multiplexed template DNA (M1-M3), wherein the DNA was first
amplified using 12 base primers that annealed 150 bases upstream
and downstream of the loci of interest. Then, three separate PCR
reactions were performed on the multiplexed template DNA, using a
first and second primer.
[0124] FIG. 19C. Quantification of the percentage of allele 2 to
allele 1 for SNP TSC310507 on original template DNA (IA) and
multiplexed template DNA (M1-M3), wherein the DNA was first
amplified using 12 base primers that annealed 150 bases upstream
and downstream of the loci of interest. Then, three separate PCR
reactions were performed on the multiplexed template DNA, using a
first and second primer.
[0125] FIG. 19D. Quantification of the percentage of allele 2 to
allele 1 for SNP TSC1335008 on original template DNA (IA) and
multiplexed template DNA (M1-M3), wherein the DNA was first
amplified using 12 base primers that annealed 150 bases upstream
and downstream of the loci of interest. Then, three separate PCR
reactions were performed on the multiplexed template DNA, using a
first and second primer.
[0126] FIG. 20. Detection of fetal DNA from plasma DNA isolated
from a pregnant female. Four SNPs wherein the maternal DNA was
homozygous were analyzed on the plasma. DNA. The maternal DNA was
homozygous for adenine at TSC0838335 (lane 1), while the plasma DNA
displayed a heterozygous pattern (lane 2). The guanine allele
represented the fetal DNA, which was clearly distinguished from the
maternal signal. Both the maternal DNA and the plasma DNA were
homozygous for adenine at TSC0418134 (lanes 3 and 4). The maternal
DNA was homozygous for guanine at TSC0129188 (lane 5), while the
plasma DNA displayed a heterozygous pattern (lane 6). The adenine
allele represented the fetal DNA. Both the maternal DNA and the
plasma DNA were homozygous for adenine at TSC0501389 (lanes 7 and
8).
DETAILED DESCRIPTION OF THE INVENTION
[0127] The present invention provides a method for detecting
genetic disorders, including but not limited to mutations,
insertions, deletions, and chromosomal abnormalities, and is
especially useful for the detection of genetic disorders of a
fetus. The method is especially useful for detection of a
translocation, addition, amplification, transversion, inversion,
aneuploidy, polyploidy, monosomy, trisomy, trisomy 21, trisomy 13,
trisomy 14, trisomy 15, trisomy 16, trisomy 18, trisomy 22,
triploidy, tetraploidy, and sex chromosome abnormalities including
but not limited to XO, XXY, XYY, and XXX. The method also provides
a non-invasive technique for determining the sequence of fetal DNA
and identifying mutations within the fetal DNA.
[0128] The invention is directed to a method for detecting
chromosomal abnormalities, the method comprising: (a) determining
the sequence of alleles of a locus of interest on a template DNA;
and (b) quantitating a ratio for the alleles at a heterozygous
locus of interest that was identified from the locus of interest of
(a), wherein said ratio indicates the presence or absence of a
chromosomal abnormality.
[0129] In another embodiment, the present invention provides a
non-invasive method for determining the sequence of a locus of
interest on fetal DNA, said method comprising: (a) obtaining a
sample from a pregnant female; (b) adding a cell lysis inhibitor,
cell membrane stabilizer or cross-linker to the sample of (a); (c)
obtaining template DNA from the sample of (b), wherein said
template DNA comprises fetal DNA and maternal DNA; and (d)
determining the sequence of a locus of interest on template
DNA.
[0130] In another embodiment, the present invention is directed to
a method for isolating DNA, said method comprising (a) obtaining a
sample containing nucleic acid; (b) adding a cell lysis inhibitor,
cell membrane stabilizer or cross-linker to sample of (a); and (c)
isolating the DNA.
[0131] In another embodiment, the present invention is directed to
a method for isolating free DNA, said method comprising (a)
obtaining a sample containing nucleic acid; (b) adding a cell lysis
inhibitor, cell membrane stabilizer or cross-linker to sample of
(a); and (c) isolating the DNA.
[0132] In another embodiment, the present invention is directed to
a method for isolating free DNA from a sample containing nucleic
acid to which a cell lysis inhibitor, cell membrane stabilizer or
cross-linker has been added, said method comprising isolating the
DNA.
[0133] In another embodiment, the present invention is directed to
a method for isolating free fetal DNA, said method comprising (a)
obtaining a sample containing nucleic acid; (b) adding a cell lysis
inhibitor, cell membrane stabilizer or cross-linker to sample of
(a); and (c) isolating the DNA. In another embodiment, the DNA is
isolated using any technique suitable in the art including but not
limited to cesium chloride gradients, gradients, sucrose gradients,
glucose gradients, centrifugation protocols, boiling, Qiagen
purification systems, QIA DNA blood purification kit, HiSpeed
Plasmid Maxi Kit, QIAfilter plasmid kit, Promega DNA purification
systems, MangeSil Paramagnetic Particle based systems, Wizard SV
technology, Wizard Genomic DNA purification kit, Amersham
purification systems, GFX Genomic Blood DNA purification kit,
Invitrogen Life Technologies Purification Systems, CONCERT
purification system, Mo Bio Laboratories purification systems,
UltraClean BloodSpin Kits, and UlraClean Blood DNA Kit.
[0134] In another embodiment, the present invention is directed to
a method for isolating free fetal DNA from a sample containing
nucleic acid to which a cell lysis inhibitor, cell membrane
stabilizer or cross-linker has been added, said method comprising
isolating the DNA. In a preferred embodiment, the free fetal DNA is
isolated from plasma or serum obtained from the blood of a pregnant
female.
[0135] In another embodiment, the DNA is isolated using techniques
and/or protocols that substantially reduce the amount of maternal
DNA in the sample including but not limited to centrifuging the
samples, with the braking power for the centrifuge set to zero (the
brake on the centrifuge is not used), transferring the supernatant
to a new tube with minimal or no disturbance of the "buffy-coat,"
and transferring only a portion of the supernatant to a new tube.
In a preferred embodiment, both acceleration power and braking
power for the centrifuge are set to zero.
[0136] In another embodiment, the DNA is isolated using techniques
and/or protocols that substantially reduce the amount of maternal
DNA in the sample including but not limited to centrifuging the
samples, with the acceleration power for the centrifuge set to
zero, transferring the supernatant to a new tube with minimal or no
disturbance of the "buffy-coat," and transferring only a portion of
the supernatant to a new tube.
[0137] In another embodiment, the "buffy-coat" is removed from the
tube prior to removal of the supernatant using any applicable
method including but not limited to using a syringe or needle to
withdraw the "buffy-coat,"
[0138] In another embodiment, the braking power for the centrifuge
is set at a percentage including but not limited to 1-5%, 5-10%,
10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%, 70-80%, 80-90%,
90-95%, 95-99% of maximum braking power.
[0139] In another embodiment, the acceleration power for the
centrifuge is set at a percentage including but not limited to
1-5%, 5-10%, 10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%,
70-80%, 80-90%, 90-95%, 95-99% of maximum acceleration power.
[0140] In another embodiment, the present invention is directed to
a composition comprising free fetal DNA and free maternal DNA,
wherein the composition comprises a relationship of free fetal DNA
to free maternal DNA including but not limited to at least about
15% free fetal DNA, at least about 20% free fetal DNA, at least
about 30% free fetal DNA, at least about 40% free fetal DNA, at
least about 50% free fetal DNA, at least about 60% free fetal DNA,
at least about 70% free fetal DNA, at least about 80% free fetal
DNA, at least about 90% free fetal DNA, at least about 91% free
fetal DNA, at least about 92% free fetal DNA, at least about 93%
free fetal DNA, at least about 94% free fetal DNA, at least about
95% free fetal DNA, at least about 96% free fetal DNA, at least
about 97% free fetal DNA, at least about 98% free fetal DNA, at
least about 99% free fetal DNA, and at least about 99.5% free fetal
DNA.
[0141] In another embodiment, the present invention is directed to
a method of using a composition comprising free fetal DNA and free
maternal DNA for prenatal diagnostics, wherein the composition
comprises a relationship of free fetal DNA to free maternal DNA
including but not limited to at least about 15% free fetal DNA, at
least about 20% free fetal DNA, at least about 30% free fetal DNA,
at least about 40% free fetal DNA, at least about 50% free fetal
DNA, at least about 60% free fetal DNA, at least about 70% free
fetal DNA, at least about 80% free fetal DNA, at least about 90%
free fetal DNA, at least about 91% free fetal DNA, at least about
92% free fetal DNA, at least about 93% free fetal DNA, at least
about 94% free fetal DNA, at least about 95% free fetal DNA, at
least about 96% free fetal DNA, at least about 97% free fetal DNA,
at least about 98% free fetal. DNA, at least about 99% free fetal
DNA, and at least about 99.5% free fetal DNA.
[0142] In another embodiment, the present invention is directed to
a composition comprising free fetal DNA and free maternal DNA,
wherein the composition comprises a relationship of free fetal DNA
to free maternal DNA including but not limited to about 13-15% free
fetal DNA, about 15-16% free fetal DNA, about 16-17% free fetal
DNA, about 17-18% free fetal DNA, about 18-19% free fetal DNA,
about 19-20% free fetal DNA, about 20-21% free fetal DNA, about
21-22% free fetal DNA, about 22-23% free fetal DNA, about 23-24%
free fetal DNA, about 24-25% free fetal DNA, about 25-35% free
fetal DNA, about 35-45% free fetal DNA, about 45-55% free fetal
DNA, about 55-65% free fetal DNA, about 65-75% free fetal DNA,
about 75-85% free fetal DNA, about 85-90% free fetal DNA, about
90-91% free fetal DNA, about 91-92% free fetal DNA, about 92-93%
free fetal DNA, about 93-94% free fetal DNA, about 94-95% free
fetal DNA, about 95-96% free fetal DNA, about 96-97% free fetal
DNA, about 97-98% free fetal DNA, about 98-99% free fetal DNA, and
about 99-99.7% free fetal DNA.
[0143] In another embodiment, the present invention is directed to
a method of using a composition comprising free fetal DNA and free
maternal DNA for prenatal diagnostics, wherein the composition
comprises a relationship of free fetal DNA to free maternal DNA
including but not limited to about 13-15% free fetal DNA, about
15-16% free fetal DNA, about 16-17% free fetal DNA, about 17-18%
free fetal DNA, about 18-19% free fetal DNA, about 19-20% free
fetal DNA, about 20-21% free fetal DNA, about 21-22% free fetal
DNA, about 22-23% free fetal DNA, about 23-24% free fetal DNA,
about 24-25% free fetal DNA, about 25-35% free fetal DNA, about
35-45% free fetal DNA, about 45-55% free fetal DNA, about 55-65%
free fetal DNA, about 65-75% free fetal DNA, about 75-85% free
fetal DNA, about 85-90% free fetal DNA, about 90-91% free fetal
DNA, about 91-92% free fetal DNA, about 92-93% free fetal DNA,
about 93-94% free fetal DNA, about 94-95% free fetal DNA, about
95-96% free fetal DNA, about 96-97% free fetal DNA, about 97-98%
free fetal DNA, about 98-99% free fetal DNA, or about 99-99.7% free
fetal DNA.
[0144] In another embodiment, the present invention is directed to
a composition comprising free fetal DNA and free maternal DNA,
wherein the composition comprises a relationship of free fetal DNA
to free maternal DNA including but not limited a maximum of 13%-15%
free fetal DNA, a maximum of 15-18% free fetal DNA, a maximum of
18-20% free fetal DNA, a maximum of 20-40% free fetal DNA, a
maximum of 40-50% free fetal DNA, a maximum of 50-60% free fetal
DNA, a maximum of 60-70% free fetal DNA, a maximum of 70-80% free
fetal DNA, a maximum of 80-90% free fetal DNA, a maximum of 90-92%
free fetal DNA, a maximum of 92-94% free fetal DNA, a maximum of
94-95% free fetal DNA, a maximum of 95-96% free fetal DNA, a
maximum of 96-97% free fetal DNA, a maximum of 97-98% free fetal
DNA, a maximum of 98-99% free fetal DNA, a maximum of 99-99.5% free
fetal DNA, and a maximum of 99.5-99.9% free fetal DNA.
[0145] In another embodiment, the present invention is directed to
a method of using a composition comprising free fetal DNA and free
maternal DNA for prenatal diagnostics, wherein the composition
comprises a relationship of free fetal DNA to free maternal DNA
including but not limited a maximum of 13%-15% free fetal DNA, a
maximum of 15-18% free fetal DNA, a maximum of 18-20% free fetal
DNA, a maximum of 20-40% free fetal DNA, a maximum of 40-50% free
fetal DNA, a maximum of 50-60% free fetal DNA, a maximum of 60-70%
free fetal DNA, a maximum of 70-80% free fetal DNA, a maximum of
80-90% free fetal DNA, a maximum of 90-92% free fetal DNA, a
maximum of 92-94% free fetal DNA, a maximum of 94-95% free fetal
DNA, a maximum of 95-96% free fetal DNA, a maximum of 96-97% free
fetal DNA, a maximum of 97-98% free fetal DNA, a maximum of 98-99%
free fetal DNA, a maximum of 99-99.5% free fetal DNA, and a maximum
of 99.5-99.9% free fetal DNA.
DNA Template
[0146] By a "locus of interest" is intended a selected region of
nucleic acid that is within a larger region of nucleic acid. A
locus of interest can include but is not limited to 1-100, 1-50,
1-20, or 1-10 nucleotides, preferably 1-6, 1-5, 1-4, 1-3, 1-2, or 1
nucleotide(s).
[0147] As used herein, an "allele" is one of several alternate
forms of a gene or non-coding regions of DNA that occupy the same
position on a chromosome. The term allele can be used to describe
DNA from any organism including but not limited to bacteria,
viruses, fungi, protozoa, molds, yeasts, plants, humans,
non-humans, animals, and archeabacteria.
[0148] For example, bacteria typically have one large strand of
DNA. The term allele with respect to bacterial DNA refers to the
form of a gene found in one cell as compared to the form of the
same gene in a different bacterial cell of the same species.
[0149] Alleles can have the identical sequence or can vary by a
single nucleotide or more than one nucleotide. With regard to
organisms that have two copies of each chromosome, if both
chromosomes have the same allele, the condition is referred to as
homozygous. If the alleles at the two chromosomes are different,
the condition is referred to as heterozygous. For example, if the
locus of interest is SNP X on chromosome 1, and the maternal
chromosome contains an adenine at SNP X (A allele) and the paternal
chromosome contains a guanine at SNP X (G allele), the individual
is heterozygous at SNP X.
[0150] As used herein, sequence means the identity of one
nucleotide or more than one contiguous nucleotides in a
polynucleotide. In the case of a single nucleotide, e.g., a SNP,
"sequence" and "identity" are used interchangeably herein.
[0151] The term "chromosomal abnormality" refers to a deviation
between the structure of the subject chromosome and a normal
homologous chromosome. The term "normal" refers to the predominate
karyotype or banding pattern found in healthy individuals of a
particular species. A chromosomal abnormality can be numerical or
structural, and includes but is not limited to aneuploidy,
polyploidy, inversion, a trisomy, a monosomy, duplication,
deletion, deletion of a part of a chromosome, addition, addition of
a part of chromosome, insertion, a fragment of a chromosome, a
region of a chromosome, chromosomal rearrangement, and
translocation. A chromosomal abnormality can be correlated with
presence of a pathological condition or with a predisposition to
develop a pathological condition. As defined herein, a single
nucleotide polymorphism ("SNP") is not a chromosomal
abnormality.
[0152] As used herein, incorporation of a nucleotide by a
polymerase is referred to as an elongation reaction or a fill-in
reaction interchangeably.
[0153] As used herein with respect to individuals, "mutant alleles"
refers to variant alleles that are associated with a disease
state.
[0154] The term "template" refers to any nucleic acid molecule that
can be used for amplification in the invention. RNA or DNA that is
not naturally double stranded can be made into double stranded DNA
so as to be used as template DNA. Any double stranded DNA or
preparation containing multiple, different double stranded DNA
molecules can be used as template DNA to amplify a locus or loci of
interest contained in the template DNA.
[0155] The template DNA can be obtained from any source including
but not limited to humans, non-humans, mammals, reptiles, cattle,
cats, dogs, goats, swine, pigs, monkeys, apes, gorillas, bulls,
cows, bears, horses, sheep, poultry, mice, rats, fish, dolphins,
whales, and sharks.
[0156] The template DNA can be from any appropriate sample
including but not limited to, nucleic acid-containing samples of
tissue, bodily fluid (for example, blood, serum, plasma, saliva,
urine, tears, peritoneal fluid, ascitic fluid, vaginal secretion,
breast fluid, breast milk, lymph fluid, cerebrospinal fluid or
mucosa secretion), umbilical cord blood, chorionic villi, amniotic
fluid, an embryo, a two-celled embryo, a four-celled embryo, an
eight-celled embryo, a 16-celled embryo, a 32-celled embryo, a
64-celled embryo, a 128-celled embryo, a 256-celled embryo, a
512-celled embryo, a 1024-celled embryo, embryonic tissues, lymph
fluid, cerebrospinal fluid, mucosa secretion, or other body
exudate, fecal matter, an individual cell or extract of the such
sources that contain the nucleic acid of the same, and subcellular
structures such as mitochondria, using protocols well established
within the art.
[0157] In one embodiment, the template DNA can be obtained from a
sample of a pregnant female.
[0158] In another embodiment, the template DNA can be obtained from
an embryo. In a preferred embodiment, the template DNA can be
obtained from a single-cell of an embryo.
[0159] In one embodiment, the template DNA is fetal DNA. Fetal DNA
can be obtained from sources including but not limited to maternal
blood, maternal serum, maternal plasma, fetal cells, umbilical cord
blood, chorionic villi, amniotic fluid, urine, saliva, cells or
tissues.
[0160] In another embodiment, a cell lysis inhibitor is added to
the sample including but not limited to formaldehyde, formaldehyde
derivatives, formalin, glutaraldehyde, glutaraldehyde derivatives,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP,
DTME, DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS or
compounds listed in Table XXIII. In another embodiment, two, three,
four, five or more than five cell lysis inhibitors can be added to
the sample. In a preferred embodiment, formalin is present in the
sample at a percentage including but not limited to 0.0001-0.03%,
0.03-0.05%, 0.05-0.08%, 0.08-0.1%, 0.1-0.3%, 0.3-0.5%, 0.5-0.3%,
0.7-0.9%, 0.9-1.2%, 1.2-1.5%, 1.5-2%, 2-3%, 3-5%, and greater than
5%. In another embodiment, any combination of cross-linker, cell
membrane stabilizer, or cell lysis inhibitor can be added to the
sample including but not limited to a cross-linker and a cell
membrane stabilizer, a cross-linker and a cell lysis inhibitor, and
a cell membrane stabilizer and a cell lysis inhibitor. More than
one cross-linker can be used with more than one cell membrane
stabilizer. More than one cross-linker can be used with more than
one cell lysis inhibitor. More than one cell membrane stabilizer
can be used with more than cell lysis inhibitor.
[0161] In another embodiment, the cell lysis inhibitor is added to
the sample such that lysis is less than about 10% of the cells. In
a preferred embodiment, the cell lysis inhibitor is added to the
sample such that lysis is less than about 5% of the cells. In a
most preferred embodiment, the cell lysis inhibitor is added to the
sample such that lysis is less than about 1% of the cells.
[0162] In another embodiment, a cell membrane stabilizer is added
to the sample such that lysis is less than about 10% of the cells.
In a preferred embodiment, the cell membrane stabilizer is added to
the sample such that lysis is less than about 5% of the cells. In a
most preferred embodiment, the cell membrane stabilizer is added to
the sample such that lysis is less than about 1% of the cells.
[0163] In another embodiment, a cross-linker is added to the sample
such that lysis is less than about 10% of the cells. In a preferred
embodiment, the cross-linker is added to the sample such that lysis
is less than about 5% of the cells. In a most preferred embodiment,
the cross-linker is added to the sample such that lysis is less
than about 1% of the cells.
[0164] In another embodiment, the cell lysis inhibitor,
cross-linker or cell membrane stabilizer is added to the sample in
an applicable time period including but not limited to 1-10
seconds, 10-30 seconds, 30-60 seconds, 1-5 minutes, 5-10 minutes,
10-20 minutes, 20-30 minutes, 30-40 minutes, 40-50 minutes, 60-90
minutes, 90-180 minutes or greater than 180 minutes after
collection of the sample. In another embodiment, the cell lysis
inhibitor, cross-linker, or cell membrane stabilizer is present in
the apparatus to which the sample is collected including but not
limited to a glass tube, a plastic tube, a circular container, an
eppendorf tube, an IV bag, or any other appropriate collection
device. In another embodiment, after the addition of the cell lysis
inhibitor, cell membrane stabilizer, or cross-linker, the sample is
left at about room temperature for the period of time to allow the
reagent to function, including but not limited to 1-5, 5-10, 10-20,
20-40, 40-60, 60-90, 90-120, 120-150, 150-180, 180-240, 240-300 or
greater than 300 minutes.
[0165] In another embodiment, the template DNA contains both
maternal DNA and fetal DNA. In a preferred embodiment, template DNA
is obtained from blood of a pregnant female. Blood is collected
using any standard technique for blood-drawing including but not
limited to venipuncture. For example, blood can be drawn from a
vein from the inside of the elbow or the back of the hand. Blood
samples can be collected from a pregnant female at any time during
fetal gestation. For example, blood samples can be collected from
human females at 1-4, 4-8, 8-12, 12-16, 16-20, 20-24, 24-28, 28-32,
32-36, 36-40, or 40-44 weeks of fetal gestation, and preferably
between 8-28 weeks of fetal gestation.
[0166] The blood sample is centrifuged to separate the plasma from
the maternal cells. The plasma and maternal cell fractions are
transferred to separate tubes and re-centrifuged. The plasma
fraction contains cell-free fetal DNA and maternal DNA. Any
standard DNA isolation technique can be used to isolate the fetal
DNA and the maternal DNA including but not limited to QIAmp DNA
Blood Midi Kit supplied by QIAGEN (Catalog number 51183).
[0167] In a preferred embodiment, blood can be collected into an
apparatus containing a magnesium chelator including but not limited
to EDTA, and is stored at 4.degree. C. Optionally, a calcium
chelator, including but not limited to EGTA, can be added.
[0168] In another embodiment, a cell lysis inhibitor is added to
the maternal blood including but not limited to formaldehyde,
formaldehyde derivatives, formalin, glutaraldehyde, glutaraldehyde
derivatives, a protein cross-linker, a nucleic acid cross-linker, a
protein and nucleic acid cross-linker, primary amine reactive
crosslinkers, sulfhydryl reactive crosslinkers, sulfhydryl addition
or disulfide reduction, carbohydrate reactive crosslinkers,
carboxyl reactive crosslinkers, photoreactive crosslinkers,
cleavable crosslinkers, AEDP, APG, BASED, BM(PEO).sub.3,
BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3, BSOCOES, DFDNB, DMA, DMP,
DMS, DPDPB, DSG, DSP, DSS, DST, DTBP, DTME, DTSSP, EGS, HBVS,
sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS, or compounds listed in Table
XXIII.
[0169] In another embodiment, an agent that stabilizes cell
membranes may be added to the maternal blood samples to reduce
maternal cell lysis including but not limited to aldehydes, urea
formaldehyde, phenol formaldehyde, DMAE (dimethylaminoethanol),
cholesterol, cholesterol derivatives, high concentrations of
magnesium, vitamin E, and vitamin E derivatives, calcium, calcium
gluconate, taurine, niacin, hydroxylamine derivatives, bimoclomol,
sucrose, astaxanthin, glucose, amitriptyline, isomer A hopane
tetral phenylacetate, isomer B hopane tetral phenylacetate,
citicoline, inositol, vitamin B, vitamin B complex, cholesterol
hemisuccinate, sorbitol, calcium, coenzyme Q, ubiquinone, vitamin
K, vitamin K complex, menaquinone, zonegran, zinc, ginkgo biloba
extract, diphenylhydantoin, perftoran, polyvinylpyrrolidone,
phosphatidylserine, tegretol, PABA, disodium cromglycate,
nedocromil sodium, phenyloin, zinc citrate, mexitil, dilantin,
sodium hyaluronate, or polaxamer 188.
[0170] In another embodiment, the template DNA is obtained from the
plasma or serum of the blood of the pregnant female. The percentage
of fetal DNA in maternal plasma is between 0.39-11.9% (Pertl, and
Bianchi, Obstetrics and Gynecology 98: 483-490 (2001)). The
majority of the DNA in the plasma sample is maternal, which makes
using the DNA for genotyping the fetus difficult. However, methods
that increase the percentage of fetal DNA in the maternal plasma
allow the sequence of the fetal DNA to be determined, and allow for
the detection of genetic disorders including mutations, insertions,
deletions, and chromosomal abnormalities. The addition of cell
lysis inhibitors, cell membrane stabilizers or cross-linkers to the
maternal blood sample can increase the relative percentage of fetal
DNA. While lysis of both maternal and fetal cells is inhibited, the
vast majority of cells are maternal, and thus by reducing the lysis
of maternal cells, there is a relative increase in the percentage
of free fetal DNA. See Example 4.
[0171] In another embodiment, any blood drawing technique, method,
protocol, or equipment that reduce the amount of cell lysis can be
used, including but not limited to a large boar needle, a shorter
length needle, a needle coating that increases laminar flow, e.g.,
teflon, a modification of the bevel of the needle to increase
laminar flow, or techniques that reduce the rate of blood flow. The
fetal cells likely are destroyed in the maternal blood by the
mother's immune system. However, it is likely that a large portion
of the maternal cell lysis occurs as a result of the blood draw or
processing of the blood sample. Thus, methods that prevent or
reduce cell lysis will reduce the amount of maternal DNA in the
sample, and increase the relative percentage of free fetal DNA.
[0172] In another embodiment, an agent that preserves or stabilizes
the structural integrity of cells can be used to reduce the amount
of cell lysis.
[0173] In another embodiment, any protocol that reduces the amount
of free maternal DNA in the maternal blood can be used prior to
obtaining the sample. In another embodiment, prior to obtaining the
sample, the pregnant female rests without physical activity for a
period of time including but not limited to 0-5, 5-10, 10-15,
15-20, 20-25, 25-30, 30-35, 35-40, 40-45, 45-50, 50-55, 55-60,
60-120, 120-180, 180-240, 240-300, 300-360, 360-420, 420-480,
480-540, 540-600, 600-660, 660-720, 720-780, 780-840, 840-900,
900-1200, 1200-1500, 1500-1800, 1800-2100, 2100-2400, 2400-2700,
2700-3000, 3000-3300, 3300-3600, 3600-3900, 3900-4200, 4200-4500,
and greater than 4500 minutes. In another embodiment, the sample is
obtained from the pregnant female after her body has reached a
relaxed state. The period of rest prior to obtaining the sample may
reduce the amount of maternal nucleic acid in the sample. In
another embodiment, the sample is obtained from the pregnant female
in the a.m., including but not limited to 4-5 am, 5-6 am, 6-7 am,
7-8 am, 8-9 am, 9-10 am, 10-11 am, and 11-12 am.
[0174] In another embodiment, the sample is obtained from the
pregnant female after she has slept for a period of time including
but not limited to 0-1, 1-2, 2-3, 3-4, 4-5, 5-6, 6-7, 7-8, 8-9,
9-10, 10-11, 11-12, or greater than 12 hours.
[0175] In another embodiment, prior to obtaining the sample, the
pregnant female exercises for a period of time followed by a period
of rest. In another embodiment, the period of exercise includes but
is not limited to 0-15, 15-30, 30-45, 45-60, 60-120, 120-240, or
greater than 240 minutes.
[0176] In another embodiment, agents that prevent the destruction
of DNA, including but not limited to a DNase inhibitor, zinc
chloride, ethylenediaminetetraacetic acid, guanidine-HCl, guanidine
isothiocyanate, N-lauroylsarcosine, and Na-dodecylsulphate, can be
added to the blood sample.
[0177] In another embodiment, fetal DNA is obtained from a fetal
cell, wherein said fetal cell can be isolated from sources
including but not limited to maternal blood, umbilical cord blood,
chorionic amniotic fluid, embryonic tissues and mucous obtained
from the cervix or vagina of the mother.
[0178] In a preferred embodiment, fetal cells are isolated from
maternal peripheral blood. An antibody specific for fetal cells can
be used to purify the fetal cells from the maternal serum (Mueller
et al., Lancet 336: 197-200 (1990); Ganshirt-Ahlert et al., Am. J.
Obstet. Gynecol. 166: 1350-1355 (1992)). Flow cytometry techniques
can also be used to enrich fetal cells (Herzenberg et al., PNAS 76:
1453-1455 (1979); Bianchi et al., PNAS 87: 3279-3283 (1990); Bruch
et al., Prenatal Diagnosis 11: 787-798 (1991)). U.S. Pat. No.
5,432,054 also describes a technique for separation of fetal
nucleated red blood cells, using a tube having a wide top and a
narrow, capillary bottom made of polyethylene. Centrifugation using
a variable speed program results in a stacking of red blood cells
in the capillary based on the density of the molecules. The density
fraction containing low density red blood cells, including fetal
red blood cells, is recovered and then differentially hemolyzed to
preferentially destroy maternal red blood cells. A density gradient
in a hypertonic medium is used to separate red blood cells, now
enriched in the fetal red blood cells from lymphocytes and ruptured
maternal cells. The use of a hypertonic solution shrinks the red
blood cells, which increases their density, and facilitate
purification from the more dense lymphocytes. After the fetal cells
have been isolated, fetal DNA can be purified using standard
techniques in the art.
[0179] The nucleic acid that is to be analyzed can be any nucleic
acid, e.g., genomic, plasmid, cosmid, yeast artificial chromosomes,
artificial or man-made DNA, including unique DNA sequences, and
also DNA that has been reverse transcribed from an RNA sample, such
as cDNA. The sequence of RNA can be determined according to the
invention if it is capable of being made into a double stranded DNA
form to be used as template DNA.
[0180] The terms "primer" and "oligonucleotide primer" are
interchangeable when used to discuss an oligonucleotide that
anneals to a template and can be used to prime the synthesis of a
copy of that template.
[0181] "Amplified" DNA is DNA that has been "copied" once or
multiple times, e.g. by polymerase chain reaction. When a large
amount of DNA is available to assay, such that a sufficient number
of copies of the locus of interest are already present in the
sample to be assayed, it may not be necessary to "amplify" the DNA
of the locus of interest into an even larger number of replicate
copies. Rather, simply "copying" the template DNA once using a set
of appropriate primers, which may contain hairpin structures that
allow the restriction enzyme recognition sites to be double
stranded, can suffice.
[0182] "Copy" as in "copied DNA" refers to DNA that has been copied
once, or DNA that has been amplified into more than one copy.
[0183] In one embodiment, the nucleic acid is amplified directly in
the original sample containing the source of nucleic acid. It is
not essential that the nucleic acid be extracted, purified or
isolated; it only needs to be provided in a form that is capable of
being amplified. Hybridization of the nucleic acid template with
primer, prior to amplification, is not required. For example,
amplification can be performed in a cell or sample lysate using
standard protocols well known in the art. DNA that is on a solid
support, in a fixed biological preparation, or otherwise in a
composition that contains non-DNA substances and that can be
amplified without first being extracted from the solid support or
fixed preparation or non-DNA substances in the composition can be
used directly, without further purification, as long as the DNA can
anneal with appropriate primers, and be copied, especially
amplified, and the copied or amplified products can be recovered
and utilized as described herein.
[0184] In a preferred embodiment, the nucleic acid is extracted,
purified or isolated from non-nucleic acid materials that are in
the original sample using methods known in the art prior to
amplification.
[0185] In another embodiment, the nucleic acid is extracted,
purified or isolated from the original sample containing the source
of nucleic acid and prior to amplification, the nucleic acid is
fragmented using any number of methods well known in the art
including but not limited to enzymatic digestion, manual shearing,
or sonication. For example, the DNA can be digested with one or
more restriction enzymes that have a recognition site, and
especially an eight base or six base pair recognition site, which
is not present in the loci of interest. Typically, DNA can be
fragmented to any desired length, including 50, 100, 250, 500,
1,000, 5,000, 10,000, 50,000 and 100,000 base pairs long. In
another embodiment, the DNA is fragmented to an average length of
about 1000 to 2000 base pairs. However, it is not necessary that
the DNA be fragmented.
[0186] Fragments of DNA that contain the loci of interest can be
purified from the fragmented DNA before amplification. Such
fragments can be purified by using primers that will be used in the
amplification (see "Primer Design" section below) as hooks to
retrieve the loci of interest, based on the ability of such primers
to anneal to the loci of interest. In a preferred embodiment,
tag-modified primers are used, such as e.g. biotinylated
primers.
[0187] By purifying the DNA fragments containing the loci of
interest, the specificity of the amplification reaction can be
improved. This will minimize amplification of nonspecific regions
of the template DNA. Purification of the DNA fragments can also
allow multiplex PCR (Polymerase Chain Reaction) or amplification of
multiple loci of interest with improved specificity.
[0188] The loci of interest that are to be sequenced can be
selected based upon sequence alone. In humans, over 1.42 million
single nucleotide polymorphisms (SNPs) have been described (Nature
409:928-933 (2001); The SNP Consortium LTD). On the average, there
is one SNP every 1.9 kb of human genome. However, the distance
between loci of interest need not be considered when selecting the
loci of interest to be sequenced according to the invention. If
more than one locus of interest on genomic DNA is being analyzed,
the selected loci of interest can be on the same chromosome or on
different chromosomes.
[0189] In a preferred embodiment, the selected loci of interest can
be clustered to a particular region on a chromosome. Multiple loci
of interest can be located within a region of DNA such that even
with any breakage or fragmentation of the DNA, the multiple loci of
interest remain linked. For example, if the DNA is obtained and by
natural forces is broken into fragments of 5 Kb, multiple loci of
interest can be selected within the 5 Kb regions. This allows each
fragment, as measured by the loci of interest within that fragment,
to serve as an experimental unit, and will reduce any possible
experimental noise of comparing loci of interest on multiple
chromosomes.
[0190] The loci of interest on a chromosome can be any distance
from each other including but not limited to 10-50, 50-100,
100-150, 150-200, 200-250, 250-500, 500-750, 750-1000, 1060-1500,
1500-2000, 2000-2500, 2500-3000, 3000-3500, 3500-4000, 4000-4500,
4500-5000, 5000-10,000 and greater than 10,000 base pairs.
[0191] In a preferred embodiment, the length of sequence that is
amplified is preferably different for each locus of interest so
that the loci of interest can be separated by size.
[0192] In fact, it is an advantage of the invention that primers
that copy an entire gene sequence need not be utilized. Rather, the
copied locus of interest is preferably only a small part of the
total gene or a small part of a non-coding region of DNA. There is
no advantage to sequencing the entire gene as this can increase
cost and delay results. Sequencing only the desired bases or loci
of interest maximizes the overall efficiency of the method because
it allows for the sequence of the maximum number of loci of
interest to be determined in the fastest amount of time and with
minimal cost.
[0193] Because a large number of sequences can be analyzed
together, the method of the invention is especially amenable to the
large-scale screening of a number of loci of interest.
[0194] Any number of loci of interest can be analyzed and
processed, especially at the same time, using the method of the
invention. The sample(s) can be analyzed to determine the sequence
at one locus of interest or at multiple loci of interest at the
same time. The loci of interest can be present on a single
chromosome or on multiple chromosomes.
[0195] Alternatively, 2, 3, 4, 5, 6, 7, 8, 9, 10-20, 20-25, 25-30,
30-35, 35-40, 40-45, 45-50, 50-100, 100-250, 250-500, 500-1,000,
1,000-2,000, 2,000-3,000, 3,000-5,000, 5,000-10,000, 10,000-50,000
or more than 50,000 loci of interest can be analyzed at the same
time when a global genetic screening is desired. Such a global
genetic screening might be desired when using the method of the
invention to provide a genetic fingerprint to identify an
individual or for SNP genotyping.
[0196] The locus of interest to be copied can be within a coding
sequence or outside of a coding sequence. Preferably, one or more
loci of interest that are to be copied are within a gene. In a
preferred embodiment, the template DNA that is copied is a locus or
loci of interest that is within a genomic coding sequence, either
intron or exon. In a highly preferred embodiment, exon DNA
sequences are copied. The loci of interest can be sites where
mutations are known to cause disease or predispose to a disease
state. The loci of interest can be sites of single nucleotide
polymorphisms. Alternatively, the loci of interest that are to be
copied can be outside of the coding sequence, for example, in a
transcriptional regulatory region, and especially a promoter,
enhancer, or repressor sequence.
Method for Determining the Sequence of a Locus of Interest
[0197] Any method that provides information on the sequence of a
nucleic acid can be used including but not limited to allele
specific PCR, PCR, gel electrophoresis, ELISA, mass spectrometry,
MALDI-TOF mass spectrometry hybridization, primer extension,
fluorescence detection, fluorescence resonance energy transfer
(FRET), fluorescence polarization, DNA sequencing, Sanger dideoxy
sequencing, DNA sequencing gels, capillary electrophoresis on an
automated DNA sequencing machine, microchannel electrophoresis,
microarray, southern blot, slot blot, dot blot, single primer
linear nucleic acid amplification, as described in U.S. Pat. No.
6,251,639, SNP-IT, GeneChips, HuSNP, BeadArray, TaqMan assay,
Invader assay, MassExtend, or MassCleave.TM. (hMC) method.
[0198] The preferred method of determining the sequence has
previously been described in U.S. application Ser. No. 10/093,618,
filed on Mar. 11, 2002, hereby incorporated by reference in its
entirety.
I. Primer Design
[0199] Published sequences, including consensus sequences, can be
used to design or select primers for use in amplification of
template DNA. The selection of sequences to be used for the
construction of primers that flank a locus of interest can be made
by examination of the sequence of the loci of interest, or
immediately thereto. The recently published sequence of the human
genome provides a source of useful consensus sequence information
from which to design primers to flank a desired human gene locus of
interest.
[0200] By "flanking" a locus of interest is meant that the
sequences of the primers are such that at least a portion of the 3'
region of one primer is complementary to the antisense strand of
the template DNA and upstream from the locus of interest site
(forward primer), and at least a portion of the 3' region of the
other primer is complementary to the sense strand of the template
DNA and downstream of the locus of interest (reverse primer). A
"primer pair" is intended a pair of forward and reverse primers.
Both primers of a primer pair anneal in a manner that allows
extension of the primers, such that the extension results in
amplifying the template DNA in the region of the locus of
interest.
[0201] Primers can be prepared by a variety of methods including
but not limited to cloning of appropriate sequences and direct
chemical synthesis using methods well known in the art (Narang et
al. Methods Enzymol. 68:90 (1979); Brown et al., Methods Enzymol.
68:109 (1979)). Primers can also be obtained from commercial
sources such as Operon Technologies, Amersham Pharmacia Biotech,
Sigma, and Life Technologies. The primers can have an identical
melting temperature. The lengths of the primers can be extended or
shortened at the 5' end or the 3' end to produce primers with
desired melting temperatures. In a preferred embodiment, one of the
primers of the prime pair is longer than the other primer. In a
preferred embodiment, the 3' annealing lengths of the primers,
within a primer pair, differ. Also, the annealing position of each
primer pair can be designed such that the sequence and length of
the primer pairs yield the desired melting temperature. The
simplest equation for determining the melting temperature of
primers smaller than 25 base pairs is the Wallace Rule
(Td=2(A+T)+4(G+C)). Computer programs can also be used to design
primers, including but not limited to Array Designer Software
(Arrayit Inc.), Oligonucleotide Probe Sequence Design Software for
Genetic Analysis (Olympus Optical Co.), NetPrimer, and DNAsis from
Hitachi Software Engineering. The TM (melting or annealing
temperature) of each primer is calculated using software programs
such as Net Primer (free web based program at
http://premierbiosoft.com/netprimer/netprlaunch/netprlaunch.html;
Internet address as of Apr. 17, 2002).
[0202] In another embodiment, the annealing temperature of the
primers can be recalculated and increased after any cycle of
amplification, including but not limited to cycle 1, 2, 3, 4, 5,
cycles 6-10, cycles 10-15, cycles 15-20, cycles 20-25, cycles
25-30, cycles 30-35, or cycles 35-40. After the initial cycles of
amplification, the 5' half of the primers is incorporated into the
products from each loci of interest, thus the TM can be
recalculated based on both the sequences of the 5' half and the 3'
half of each primer.
[0203] For example, in FIG. 1B, the first cycle of amplification is
performed at about the melting temperature of the 3' region, which
anneals to the template DNA, of the second primer (region "c"),
which is 13 bases. After the first cycle, the annealing temperature
can be raised to TM2, which is about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which is depicted as region "b." The second primer cannot bind to
the original template DNA because it only anneals to 13 bases in
the original DNA template, and TM2 is about the melting temperature
of approximately 20 bases, which is the 3' annealing region of the
first primer (FIG. 1C). However, the first primer can bind to the
DNA that was copied in the first cycle of the reaction. In the
third cycle, the annealing temperature is raised to TM3, which is
about the melting temperature of the entire sequence of the second
primer, which is depicted as regions "c" and "d." The DNA template
produced from the second cycle of PCR contains both regions c' and
d', and therefore, the second primer can anneal and extend at TM3
(FIG. 1D). The remaining cycles are performed at TM3. The entire
sequence of the first primer (a+b') can anneal to the template from
the third cycle of PCR, and extend (FIG. 1E). Increasing the
annealing temperature will decrease non-specific binding and
increase the specificity of the reaction, which is especially
useful if amplifying a locus of interest from human genomic DNA,
which is about 3.times.10.sup.9 base pairs long.
[0204] As used herein, the term "about" with regard to annealing
temperatures is used to encompass temperatures within 10 degrees
celsius of the stated temperatures.
[0205] In one embodiment, one primer pair is used for each locus of
interest. However, multiple primer pairs can be used for each locus
of interest.
[0206] In one embodiment, primers are designed such that one or
both primers of the primer pair contain sequence in the 5' region
for one or more restriction endonucleases (restriction enzyme).
[0207] As used herein, with regard to the position at which
restriction enzymes digest DNA, the "sense" strand is the strand
reading 5' to 3' in the direction in which the restriction enzyme
cuts. For example, BsmF I recognizes the following sequences:
TABLE-US-00001 5' GGGAC(N).sub.10 3' (SEQ ID NO: 1) 5'
(N).sub.14GTCCC 3' (SEQ ID NO: 2) 3' CCCTG(N).sub.14 5' (SEQ ID NO:
2) 3'(N).sub.10CAGGG 5' (SEQ ID NO: 1)
[0208] The sense strand is the strand containing the "GGGAC"
sequence as it reads 5' to 3' in the direction that the restriction
enzyme cuts.
[0209] As used herein, with regard to the position at which
restriction enzymes digest DNA, the "antisense" strand is the
strand reading 3' to 5' in the direction in which the restriction
enzyme cuts.
[0210] In another embodiment, one of the primers in a primer pair
is designed such that it contains a restriction enzyme recognition
site for a restriction enzyme that cuts "n" nucleotides away from
the recognition site, and produces a recessed 3' end and a 5'
overhang that contains the locus of interest (herein referred to as
a "second primer"). "N" is a distance from the recognition site to
the site of the cut by the restriction enzyme. In other words, the
second primer of a primer pair contains a recognition site for a
restriction enzyme that does not cut DNA at the recognition site
but cuts "n" nucleotides away from the recognition site. For
example, if the recognition sequence is for the restriction enzyme
BceA I, the enzyme will cut ten (10) nucleotides from the
recognition site on the sense strand, and twelve (12) nucleotides
away from the recognition site on the antisense strand.
[0211] The 3' region and preferably, the 3' half, of the primers is
designed to anneal to a sequence that flanks the loci of interest
(FIG. 1A). The second primer can anneal any distance from the locus
of interest provided that digestion with the restriction enzyme
that recognizes the restriction enzyme recognition site on this
primer generates a 5' overhang that contains the locus of interest.
The 5' overhangs can be of any size, including but not limited to
1, 2, 3, 4, 5, 6, 7, 8, and more than 8 bases.
[0212] In a preferred embodiment, the 3' end of the primer that
anneals closer to the locus of interest (second primer) can anneal
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or more than 14
bases from the locus of interest or at the locus of interest.
[0213] In a preferred embodiment, the second primer is designed to
anneal closer to the locus of interest than the other primer of a
primer pair (the other primer is herein referred to as a "first
primer"). The second primer can be a forward or reverse primer and
the first primer can be a reverse or forward primer, respectively.
Whether the first or second primer should be the forward or reverse
primer can be determined by which design will provide better
sequencing results.
[0214] For example, the primer that anneals closer to the locus of
interest can contain a recognition site for the restriction enzyme
BsmF I, which cuts ten (10) nucleotides from the recognition site
on the sense strand, and fourteen (14) nucleotides from the
recognition site on the antisense strand. In this case, the primer
can be designed so that the restriction enzyme recognition site is
13 bases, 12 bases, 10 bases or 11 bases from the locus of
interest. If the recognition site is 13 bases from the locus of
interest, digestion with BsmF I will generate a 5' overhang (RXXX),
wherein the locus of interest (R) is the first nucleotide in the
overhang (reading 3' to 5'), and X is any nucleotide. If the
recognition site is 12 bases from the locus of interest, digestion
with BsmF I will generate a 5' overhang (XRXX), wherein the locus
of interest (R) is the second nucleotide in the overhang (reading
3' to 5'). If the recognition site is 11 bases from the locus of
interest, digestion with BsmF I will generate a 5' overhang (XXRX),
wherein the locus of interest (R) is the third nucleotide in the
overhang (reading 3' to 5'). The distance between the restriction
enzyme recognition site and the locus of interest should be
designed so that digestion with the restriction enzyme generates a
5' overhang, which contains the locus of interest. The effective
distance between the recognition site and the locus of interest
will vary depending on the choice of restriction enzyme.
[0215] In another embodiment, the primer that anneals closer to the
locus of interest site, relative to the other primer, can be
designed so that the restriction enzyme that generates the 5'
overhang, which contains the locus of interest, will see the same
sequence at the cut site, independent of the nucleotide at the
locus of interest site. For example, if the primer that anneals
closer to the locus of interest is designed so that the recognition
site for the restriction enzyme BsmF I (5' GGGAC 3') is thirteen
bases from the locus of interest, the restriction enzyme will cut
the antisense strand one base from the locus of interest. The
nucleotide at the locus of interest is adjacent to the cut site,
and may vary from DNA molecule to DNA molecule. If it is desired
that the nucleotides adjacent to the cut site be identical, the
primer can be designed so that the restriction enzyme recognition
site for BsmF I is twelve bases away from the locus of interest
site. Digestion with BsmF I will generate a 5' overhang, wherein
the locus of interest site is in the second position of the
overhang (reading 3' to 5') and is no longer adjacent to the cut
site. Designing the primer so that the restriction enzyme
recognition site is twelve (12) bases from the locus of interest
site allows the nucleotides adjacent to the cut site to be the
same, independent of the nucleotide at the locus of interest. Also,
primers that have been designed so that the restriction enzyme
recognition site, BsmF I, is eleven (11) or ten (10) bases from the
locus of interest site will allow the nucleotides adjacent to the
cut site to be the same, independent of the nucleotide at the locus
of interest. Similar strategies of primer design can be employed
with other restriction enzymes so that the nucleotides adjacent to
the cut site will be the same, independent of the nucleotide at the
loci of interest.
[0216] The 3' end of the first primer (either the forward or the
reverse) can be designed to anneal at a chosen distance from the
locus of interest. Preferably, for example, this distance is
between 1-10, 10-25, 25-50, 50-75, 75-100, 100-150, 150-200,
200-250, 250-300, 300-350, 350-400, 400-450, 450-500, 500-550,
550-600, 600-650, 650-700, 700-750, 750-800, 800-850, 850-900,
900-950, 950-1000 and greater than 1000 bases away from the locus
of interest. The annealing sites of the first primers are chosen
such that each successive upstream primer is further and further
away from its respective downstream primer.
[0217] For example, if at locus of interest 1 the 3' ends of the
first and second primers are Z bases apart, then at locus of
interest 2, the 3' ends of the upstream and downstream primers are
Z+K bases apart, where K=1, 2, 3, 4, 5-10, 10-20, 20-30, 30-40,
40-50, 50-60, 60-70, 70-80, 80-90, 90-100, 100-200, 200-300,
300-400, 400-500, 500-600, 600-700, 700-800, 800-900, 900-1000, or
greater than 1000 bases (FIG. 2). The purpose of making the first
primers further and further apart from their respective second
primers is so that the PCR products of all the loci of interest
differ in size and can be separated, e.g., on a sequencing gel.
This allows for multiplexing by pooling the PCR products in later
steps.
[0218] In one embodiment, the 5' region of the first or second
primer can have a recognition site for any type of restriction
enzyme. In a preferred embodiment, the 5' region of the first
and/or second primer has at least one restriction enzyme
recognition site that is different from the restriction enzyme
recognition site that is used to generate the 5' overhang, which
contains the locus of interest.
[0219] In one embodiment, the 5' region of the first primer can
have a recognition site for any type of restriction enzyme. In a
preferred embodiment, the first primer has at least one restriction
enzyme recognition site that is different from the restriction
enzyme recognition site in the second primer. In another preferred
embodiment, the first primer anneals further away from the locus of
interest than the second primer.
[0220] In a preferred embodiment, the second primer contains a
restriction enzyme recognition sequence for a Type IIS restriction
enzyme including but not limited to BceA I and BsmF I, which
produce a two base 5' overhang and a four base 5' overhang,
respectively. Restriction enzymes that are Type IIS are preferred
because they recognize asymmetric base sequences (not palindromic
like the orthodox Type II enzymes). Type IIS restriction enzymes
cleave DNA at a specified position that is outside of the
recognition site, typically up to 20 base pairs outside of the
recognition site. These properties make Type IIS restriction
enzymes, and the recognition sites thereof, especially useful in
the method of the invention. Preferably, the Type IIS restriction
enzymes used in this method leave a 5' overhang and a recessed
3'.
[0221] A wide variety of Type IIS restriction enzymes are known and
such enzymes have been isolated from bacteria, phage,
archeabacteria and viruses of eukaryotic algae and are commercially
available (Promega, Madison Wis.; New England Biolabs, Beverly,
Mass.; Szybalski W. et al., Gene 100:13-26, 1991). Examples of Type
IIS restriction enzymes that would be useful in the method of the
invention include, but are not limited to enzymes such as those
listed in Table I.
TABLE-US-00002 TABLE I Recognition/ Enzyme- Source Cleavage Site
Supplier Alw I - Acinetobacter lwoffii GGATC(4/5) NE Bioiabs Alw26
I - Acinetobacter lwojfi GTCTC(1/5) Promega Bbs I - Bacillus
laterosporus GAAGAC(2/6) NE Bioiabs Bbv I - Bacillus brevis
GCAGC(8/12) NE Bioiabs BceA I - Bacillus cereus 1315 IACGGC(12/14)
NE Bioiabs Bmr I - Bacillus megalerium CTGGG(5/4) NE Bioiabs Bsa I
- Bacillus stearothermophilus 6-55 GGTCTC(1/5) NE Bioiabs Bst71 I -
Bacillus stearothermophilus 71 GCAGC(8/12) Promega BsmA I -
Bacillus stearothermophilus A664 GTCTC(1/5) NE Bioiabs BsmB I
-Bacillus stearothermophilus B61 CGTCTC(1/5) NE Bioiabs BsmF I -
Bacillus stearothermophilus F GGGAC(10/14) NE Bioiabs BspM I -
Bacillus species M ACCTGC(4/8) NE Bioiabs Ear I - Enterobacter
aerogenes CTCTTC(1/4) NE Biolabs Fau I - Flavobacterium aquatile
CCCGC(4/6) NE Biolabs Fok I - Flavobacterium okeonokoites
GGATG(9/13) NE Biolabs Hga I - Haemophilus gallinarum GACGC(5/10)
NE Biolabs Ple I - Pseudomonas lemoignei GAGTC(4/5) NE Biolabs Sap
I - Saccharopolyspora species GCTCTTC(1/4) NE Biolabs SfaN I -
Streptococcus faecalis ND547 GCATC(5/9) NE Biolabs Sth132 I -
Streptococcus thermophilus STI32 CCCG(4/8) No commercial supplier
(Gene 195: 201-206 (1997))
[0222] In one embodiment, a primer pair has sequence at the 5'
region of each of the primers that provides a restriction enzyme
recognition site that is unique for one restriction enzyme.
[0223] In another embodiment, a primer pair has sequence at the 5'
region of each of the primers that provide a restriction site that
is recognized by more than one restriction enzyme, and especially
for more than one Type IIS restriction enzyme. For example, certain
consensus sequences can be recognized by more than one enzyme. For
example, BsgI, Eco571 and BpmI all recognize the consensus
(G/C)TGnAG and cleave 16 by away on the antisense strand and 14 by
away on the sense strand. A primer that provides such a consensus
sequence would result in a product that has a site that can be
recognized by any of the restriction enzymes BsgI, Eco571 and
BpmI.
[0224] Other restriction enzymes that cut DNA at a distance from
the recognition site, and produce a recessed 3' end and a 5'
overhang include Type III restriction enzymes.
[0225] For example, the restriction enzyme EcoP15I recognizes the
sequence 5' CAGCAG 3' and cleaves 25 bases downstream on the sense
strand and 27 bases on the antisense strand. It will be further
appreciated by a person of ordinary skill in the art that new
restriction enzymes are continually being discovered and can
readily be adopted for use in the subject invention.
[0226] In another embodiment, the second primer can contain a
portion of the recognition sequence for a restriction enzyme,
wherein the full recognition site for the restriction enzyme is
generated upon amplification of the template DNA such that
digestion with the restriction enzyme generates a 5' overhang
containing the locus of interest. For example, the recognition site
for BsmF I is 5' GGGACN.sub.10.sup..dwnarw. 3' (SEQ ID NO: 1). The
3' region, which anneals to the template DNA, of the second primer
can end with the nucleotides "GGG," which do not have to be
complementary with the template DNA. If the 3' annealing region is
about 10-20 bases, even if the last three bases do not anneal, the
primer will extend and, generate a BsmF I site.
TABLE-US-00003 Second primer: (SEQ ID NO: 3) 5'
GGAAATTCCATGATGCGTGGG.fwdarw. Template DNA (SEQ ID NO: 27) 3'
CCTTTAAGGTACTACGCAN.sub.1N.sub.2N.sub.3TG 5' (SEQ ID NO: 4) 5'
GGAAATTCCATGATGCCTN.sub.1,N.sub.2,N.sub.3,AC 3'
[0227] The second primer can be designed to anneal to the template
DNA, wherein the next two bases of the template DNA are thymidine
and guanine, such that an adenosine and cytosine are incorporated
into the primer forming a recognition site for BsmF I, 5'
GGGACN.sub.10.sup..dwnarw. 3' (SEQ ID NO: 1). The second primer can
be designed to anneal in such a manner that digestion with BsmF I
generates a 5' overhang containing the locus of interest.
[0228] In another embodiment, the second primer can contain an
entire or full recognition site for a restriction enzyme or a
portion of a recognition site, which generates a full recognition
site upon primer-dependent replication of the template DNA such
that digestion with a restriction enzyme that cuts at the
recognition site and generates a 5' overhang that contains the
locus of interest. For example, the restriction enzyme BsaJ I binds
the following recognition site: 5' C.sup..dwnarw.CN.sub.1N.sub.2GG
3'. The second primer can be designed such that the 3' region,
which anneals to the template DNA of the primer ends with "CC", the
SNP of interest is represented by "N.sub.1", and the template
sequence downstream of the SNP is "N.sub.2GG."
TABLE-US-00004 Second primer: (SEQ ID NO: 5) 5'
GGAAATTCCATGATGCGTACC.fwdarw. Template DNA (SEQ ID NO: 28) 3'
CCTTTAAGGTACTACGCATGGN.sub.1N.sub.2CC 5' (SEQ ID NO: 6) 5'
GGAAATTCCATGATGCCTACCN.sub.1,N.sub.2,GG 3'
[0229] After digestion with BsaJ I, a 5' overhang of the following
sequence would be generated:
TABLE-US-00005 5' C 3' 3' GGN.sub.1N.sub.2CC 5'
[0230] If the nucleotide guanine is not reported at the locus of
interest, the 3' recessed end can be filled in with unlabeled
cytosine, which is complementary to the first nucleotide in the
overhang. After removing the excess cytosine, labeled ddNTPs can be
used to fill in the next nucleotide, N.sub.1, which represents the
locus of interest. Other restriction enzymes can be used including
but not limited to BssK I (5' .sup..dwnarw.CCNGG 3'), Dde I (5'
C.sup..dwnarw.TNAG 3'), EcoN I (5' CCTNN.sup..dwnarw.NNNAGG 3' (SEQ
ID NO: 7)), Fnu4H I (5' GC.sup..dwnarw.NGC 3'), Hinf I (5'
G.sup..dwnarw.ANTC 3') PflF I (5' GACN.sup..dwnarw.NNGTC 3'), Sau96
I (5' G.sup..dwnarw.GNCC 3'), ScrF I (5' CC.sup..dwnarw.NGG 3'),
and Tth1 11 I (5' GACN.sup..dwnarw.NNGTC 3').
[0231] It is not necessary that the 3' region, which anneals to the
template DNA, of the second primer be 100% complementary to the
template DNA. For example, the last 1, 2, or 3 nucleotides of the
3' end of the second primer can be mismatches with the template
DNA. The region of the primer that anneals to the template DNA will
target the primer, and allow the primer to extend. Even if the last
two nucleotides are not complementary to the template DNA, the
primer will extend and generate a restriction enzyme recognition
site. For example, the last two nucleotides in the second primer
are "CC." The second primer anneals to the template DNA, and allows
extension even if "CC" is not complementary to the nucleotides Na,
and Nb, on the template DNA.
TABLE-US-00006 Second primer: (SEQ ID NO: 5) 5'
GGAAATTCCATGATGCGTACC.fwdarw. Template DNA (SEQ ID NO: 29) 3'
CCTTTAAGGTACTACGCATN.sub.a,N.sub.b,N.sub.1,N.sub.2,CC 5' (SEQ ID
NO: 8) 5' GGAAATTCCATGATGCCTAN.sub.aN.sub.bN.sub.1N.sub.2GG 3'
[0232] After digestion with BsaJ I, a 5' overhang of the following
sequence would be generated:
TABLE-US-00007 5' C 3' 3' GGN.sub.1N.sub.2CC 5'
[0233] If the nucleotide guanine is not reported at the locus of
interest, the 5' overhang can be filled in with unlabeled cytosine.
The excess cytosine can be rinsed away, and filled in with labeled
ddNTPs. The first nucleotide incorporated (N.sub.1') corresponds to
the locus of interest. If guanine is reported at the locus of
interest, the loci of interest can be filled in with unlabeled
cytosine and a nucleotide downstream of the locus of interest can
be detected. For example, assume N.sub.2 is adenine. If the locus
of interest is guanine, unlabeled cytosine can be used in the fill
in reaction. After removing the cytosine, a fill in reaction with
labeled thymidine can be used. The labeled thymidine will be
incorporated only if the locus of interest was a guanine. Thus, the
sequence of the locus of interest can be determined by detecting a
nucleotide downstream of the locus of interest.
[0234] In another embodiment, the first and second primers contain
a portion of a recognition sequence for a restriction enzyme,
wherein the full recognition site for the restriction enzyme is
generated upon amplification of the template DNA such that
digestion with the restriction enzyme generates a 5' overhang
containing the locus of interest. The recognition site for any
restriction enzyme that contains one or more than one variable
nucleotide can be generated including but not limited to the
restriction enzymes BssK I (5'.sup..dwnarw.CCNGG 3'), Dde I
(5'C.sup..dwnarw.TNAG 3'), Econ I (5'CCTNN.sup..dwnarw.NNNAGG 3'
(SEQ ID NO: 7)), Fnu4H I (5'GC.sup..dwnarw.NGC 3'), Hinf I
(5'G.sup..dwnarw.ANTC 3'), PflF I (5' GACN.sup..dwnarw.NNGTC 3'),
Sau96 I (5' G.sup..dwnarw.GNCC 3'), ScrF I (5' CC.sup..dwnarw.NGG
3'), and Tth1 11 I (5' GACN.sup..dwnarw.NNGTC 3').
[0235] In a preferred embodiment, the 3' regions of the first and
second primers contain the partial sequence for a restriction
enzyme, wherein the partial sequence contains 1, 2, 3, 4 or more
than 4 mismatches with the template DNA; these mismatches create
the restriction enzyme recognition site. The number of mismatches
that can be tolerated at the 3' end depends on the length of the
primer. For example, if the locus of interest is represented by
N.sub.1, a first primer can be designed to be complementary to the
template DNA, depicted below as region "a." The 3' region of the
first primer ends with "CC," which is not complementary to the
template DNA. The second primer is designed to be complementary to
the template DNA, which is depicted below as region "b'". The 3'
region of the second primer ends with "CC," which is not
complementary to the template DNA.
##STR00001##
[0236] After one round of amplification the following products
would be generated:
##STR00002##
[0237] In cycle two, the primers can anneal to the templates that
were generated from the first cycle of PCR:
##STR00003##
[0238] After cycle two of PCR, the following products would be
generated:
##STR00004##
[0239] The restriction enzyme recognition site for BsaJ I is
generated, and after digestion with BsaJ 1, a 5' overhang
containing the locus of interest is created. The locus of interest
can be detected as described in detail below.
[0240] In another embodiment, a primer pair has sequence at the 5'
region of each of the primers that provides two or more restriction
sites that are recognized by two or more restriction enzymes.
[0241] In a most preferred embodiment, a primer pair has different
restriction enzyme recognition sites at the 5' regions, especially
5' ends, such that a different restriction enzyme is required to
cleave away any undesired sequences. For example, the first primer
for locus of interest "A" can contain sequence recognized by a
restriction enzyme, "X," which can be any type of restriction
enzyme, and the second primer for locus of interest "A," which
anneals closer to the locus of interest, can contain sequence for a
restriction enzyme, "Y," which is a Type IIS restriction enzyme
that cuts "n" nucleotides away and leaves a 5' overhang and a
recessed 3' end. The 5' overhang contains the locus of interest.
After binding the amplified DNA to streptavidin coated wells, one
can digest with enzyme "Y," rinse, then fill in with labeled
nucleotides and rinse, and then digest with restriction enzyme "X,"
which will release the DNA fragment containing the locus of
interest from the solid matrix. The locus of interest can be
analyzed by detecting the labeled nucleotide that was "filled in"
at the locus of interest, e.g. SNP site.
[0242] In another embodiment, the second primers for the different
loci of interest that are being amplified according to the
invention contain recognition sequence in the 5' regions for the
same restriction enzyme and likewise all the first primers also
contain the same restriction enzyme recognition site, which is a
different enzyme from the enzyme that recognizes the second
primers.
[0243] In another embodiment, the second primers for the multiple
loci of interest that are being amplified according to the
invention contain restriction enzyme recognition sequences in the
5' regions for different restriction enzymes.
[0244] In another embodiment, the first primers for the multiple
loci of interest that are being amplified according to the
invention contain restriction enzyme recognition sequences in the
5' regions for different restriction enzymes. Multiple restriction
enzyme sequences provide an opportunity to influence the order in
which pooled loci of interest are released from the solid support.
For example, if 50 loci of interest are amplified, the first
primers can have a tag at the extreme 5' end to aid in purification
and a restriction enzyme recognition site, and the second primers
can contain a recognition site for a type IIS restriction enzyme.
For example, several of the first primers can have a restriction
enzyme recognition site for EcoR I, other first primers can have a
recognition site for Pst I, and still other first primers can have
a recognition site for BamH I. After amplification, the loci of
interest can be bound to a solid support with the aid of the tag on
the first primers. By performing the restriction digests one
restriction enzyme at a time, one can serially release the
amplified loci of interest. If the first digest is performed with
EcoR I, the loci of interest amplified with the first primers
containing the recognition site for EcoR I will be released, and
collected while the other loci of interest remain bound to the
solid support. The amplified loci of interest can be selectively
released from the solid support by digesting with one restriction
enzyme at a time. The use of different restriction enzyme
recognition sites in the first primers allows a larger number of
loci of interest to be amplified in a single reaction tube.
[0245] In a preferred embodiment, any region 5' of the restriction
enzyme digestion site of each primer can be modified with a
functional group that provides for fragment manipulation,
processing, identification, and/or purification. Examples of such
functional groups, or tags, include but are not limited to biotin,
derivatives of biotin, carbohydrates, haptens, dyes, radioactive
molecules, antibodies, and fragments of antibodies, peptides, and
immunogenic molecules.
[0246] In another embodiment, the template DNA can be replicated
once, without being amplified beyond a single round of replication.
This is useful when there is a large amount of the DNA available
for analysis such that a large number of copies of the loci of
interest are already present in the sample, and further copies are
not needed. In this embodiment, the primers are preferably designed
to contain a "hairpin" structure in the 5' region, such that the
sequence doubles back and anneals to a sequence internal to itself
in a complementary manner. When the template DNA is replicated only
once, the DNA sequence comprising the recognition site would be
single-stranded if not for the "hairpin" structure. However, in the
presence of the hairpin structure, that region is effectively
double stranded, thus providing a double stranded substrate for
activity by restriction enzymes.
[0247] To the extent that the reaction conditions are compatible,
all the primer pairs to analyze a locus or loci of interest of DNA
can be mixed together for use in the method of the invention. In a
preferred embodiment, all primer pairs are mixed with the template
DNA in a single reaction vessel. Such a reaction vessel can be, for
example, a reaction tube, or a well of a microtiter plate.
[0248] Alternatively, to avoid competition for nucleotides and to
minimize primer dimers and difficulties with annealing temperatures
for primers, each locus of interest or small groups of loci of
interest can be amplified in separate reaction tubes or wells, and
the products later pooled if desired. For example, the separate
reactions can be pooled into a single reaction vessel before
digestion with the restriction enzyme that generates a 5' overhang,
which contains the locus of interest or SNP site, and a 3' recessed
end. Preferably, the primers of each primer pair are provided in
equimolar amounts. Also, especially preferably, each of the
different primer pairs is provided in equimolar amounts relative to
the other pairs that are being used.
[0249] In another embodiment, combinations of primer pairs that
allow efficient amplification of their respective loci of interest
can be used (see e.g. FIG. 2). Such combinations can be determined
prior to use in the method of the invention. Multi-well plates and
PCR machines can be used to select primer pairs that work
efficiently with one another. For example, gradient PCR machines,
such as the Eppendorf Mastercycler.RTM. gradient PCR machine, can
be used to select the optimal annealing temperature for each primer
pair. Primer pairs that have similar properties can be used
together in a single reaction tube.
[0250] In another embodiment, a multi-sample container including
but not limited to a 96-well or more plate can be used to amplify a
single locus of interest with the same primer pairs from multiple
template DNA samples with optimal PCR conditions for that locus of
interest. Alternatively, a separate multi-sample container can be
used for amplification of each locus of interest and the products
for each template DNA sample later pooled. For example, gene A from
96 different DNA samples can be amplified in microtiter plate 1,
gene B from 96 different DNA samples can be amplified in microtiter
plate 2, etc., and then the amplification products can be
pooled.
[0251] The result of amplifying multiple loci of interest is a
preparation that contains representative PCR products having the
sequence of each locus of interest. For example, if DNA from only
one individual is used as the template DNA and if hundreds of
disease-related loci of interest were amplified from the template
DNA, the amplified DNA would be a mixture of small, PCR products
from each of the loci of interest. Such a preparation could be
further analyzed at that time to determine the sequence at each
locus of interest or at only some loci of interest. Additionally,
the preparation could be stored in a manner that preserves the DNA
and can be analyzed at a later time. Information contained in the
amplified DNA can be revealed by any suitable method including but
not limited to fluorescence detection, sequencing, gel
electrophoresis, and mass spectrometry (see "Detection of
Incorporated Nucleotide" section below).
II. Amplification of Loci of Interest
[0252] The template DNA can be amplified using any suitable method
known in the art including but not limited to PCR (polymerase chain
reaction), 3SR (self-sustained sequence reaction), LCR (ligase
chain reaction), RACE-PCR (rapid amplification of cDNA ends), PLCR
(a combination of polymerase chain reaction and ligase chain
reaction), Q-beta phage amplification (Shah et al., J. Medical
Micro. 33: 1435-41 (1995)), SDA (strand displacement
amplification), SOE-PCR (splice overlap extension PCR), and the
like. These methods can be used to design variations of the
releasable primer mediated cyclic amplification reaction explicitly
described in this application. In the most preferred embodiment,
the template DNA is amplified using PCR (PCR: A Practical Approach,
M. J. McPherson, et al., IRL Press (1991); PCR Protocols: A Guide
to Methods and Applications, Innis, et al., Academic Press (1990);
and PCR Technology: Principals and Applications of DNA
Amplification, H. A. Erlich, Stockton Press (1989)). PCR is also
described in numerous U.S. patents, including U.S. Pat. Nos.
4,683,195; 4,683,202; 4,800,159; 4,965,188; 4,889,818; 5,075,216;
5,079,352; 5,104,792, 5,023,171; 5,091,310; and 5,066,584.
[0253] The components of a typical PCR reaction include but are not
limited to a template DNA, primers, a reaction buffer (dependent on
choice of polymerase), dNTPs (dATP, dTTP, dGTP, and dCTP) and a DNA
polymerase. Suitable PCR primers can be designed and prepared as
discussed above (see "Primer Design" section above). Briefly, the
reaction is heated to 95.degree. C. for 2 min. to separate the
strands of the template DNA, the reaction is cooled to an
appropriate temperature (determined by calculating the annealing
temperature of designed primers) to allow primers to anneal to the
template DNA, and heated to 72.degree. C. for two minutes to allow
extension.
[0254] In a preferred embodiment, the annealing temperature is
increased in each of the first three cycles of amplification to
reduce non-specific amplification. See also Example 1, below. The
TM1 of the first cycle of PCR is about the melting temperature of
the 3' region of the second primer that anneals to the template
DNA. The annealing temperature can be raised in cycles 2-10,
preferably in cycle 2, to TM2, which is about the melting
temperature of the 3' region, which anneals to the template DNA, of
the first primer. If the annealing temperature is raised in cycle
2, the annealing temperature remains about the same until the next
increase in annealing temperature. Finally, in any cycle subsequent
to the cycle in which the annealing temperature was increased to
TM2, preferably cycle 3, the annealing temperature is raised to
TM3, which is about the melting temperature of the entire second
primer. After the third cycle, the annealing temperature for the
remaining cycles can be at about TM3 or can be further increased.
In this example, the annealing temperature is increased in cycles 2
and 3. However, the annealing temperature can be increased from a
low annealing temperature in cycle 1 to a high annealing
temperature in cycle 2 without any further increases in temperature
or the annealing temperature can progressively change from a low
annealing temperature to a high annealing temperature in any number
of incremental steps. For example, the annealing temperature can be
changed in cycles 2, 3, 4, 5, 6, etc.
[0255] After annealing, the temperature in each cycle is increased
to an "extension" temperature to allow the primers to "extend" and
then following extension the temperature in each cycle is increased
to the denaturization temperature. For PCR products less than 500
base pairs in size, one can eliminate the extension step in each
cycle and just have denaturization and annealing steps. A typical
PCR reaction consists of 25-45 cycles of denaturation, annealing
and extension as described above. However, as previously noted, one
cycle of amplification (one copy) can be sufficient for practicing
the invention.
[0256] In another embodiment, multiple sets of primers wherein a
primer set comprises a forward primer and a reverser primer, can be
used to amplify the template DNA for 1-5, 5-10, 10-15, 15-20 or
more than 20 cycles, and then the amplified product is further
amplified in a reaction with a single primer set or a subset of the
multiple primer sets. In a preferred embodiment, a low
concentration of each primer set is used to minimize primer-dimer
formation. A low concentration of starting DNA can be amplified
using multiple primer sets. Any number of primer sets can be used
in the first amplification reaction including but not limiting to
1-10, 10-20, 20-30, 30-40, 40-50, 50-60, 60-70, 70-80, 80-90,
90-100, 100-150, 150-200, 200-250, 250-300, 300-350, 350-400,
400-450, 450-500, 500-1000, and greater than 1000. In another
embodiment, the amplified product is amplified in a second reaction
with a single primer set. In another embodiment, the amplified
product is further amplified with a subset of the multiple primer
pairs including but not limited to 2-10, 10-20, 20-30, 30-40,
40-50, 50-60, 60-70, 70-80, 80-90, 90-100, 100-150, 150-200,
200-250, and more than 250.
[0257] The multiple primer sets will amplify the loci of interest,
such that a minimal amount of template DNA is not limiting for the
number of loci that can be detected. For example, if template DNA
is isolated from a single cell or the template DNA is obtained from
a pregnant female, which comprises both maternal template DNA and
fetal template DNA, low concentrations of each primer set can be
used in a first amplification reaction to amplify the loci of
interest. The low concentration of primers reduces the formation of
primer-dimer and increases the probability that the primers will
anneal to the template DNA and allow the polymerase to extend. The
optimal number of cycles performed with the multiple primer sets is
determined by the concentration of the primers. Following the first
amplification reaction, additional primers can be added to further
amplify the loci of interest. Additional amounts of each primer set
can be added and further amplified in a single reaction.
Alternatively, the amplified product can be further amplified using
a single primer set in each reaction or a subset of the multiple
primers sets. For example, if 150 primer sets were used in the
first amplification reaction, subsets of 10 primer sets can be used
to further amplify the product from the first reaction.
[0258] Any DNA polymerase that catalyzes primer extension can be
used including but not limited to E. coli DNA polymerase, Klenow
fragment of E. coli DNA polymerase 1, T7 DNA polymerase, T4 DNA
polymerase, Taq polymerase, Pfu DNA polymerase, Vent. DNA
polymerase, bacteriophage 29, REDTaq.TM. Genomic DNA polymerase, or
sequenase. Preferably, a thermostable DNA polymerase is used. A
"hot start" PCR can also be performed wherein the reaction is
heated to 95.degree. C. for two minutes prior to addition of the
polymerase or the polymerase can be kept inactive until the first
heating step in cycle 1. "Hot start" PCR can be used to minimize
nonspecific amplification. Any number of PCR cycles can be used to
amplify the DNA, including but not limited to 2, 5, 10, 15, 20, 25,
30, 35, 40, or 45 cycles. In a most preferred embodiment, the
number of PCR cycles performed is such that equimolar amounts of
each loci of interest are produced.
III. Purification of Amplified DNA
[0259] Purification of the amplified DNA is not necessary for
practicing the invention. However, in one embodiment, if
purification is preferred, the 5' end of the primer (first or
second primer) can be modified with a tag that facilitates
purification of the PCR products. In a preferred embodiment, the
first primer is modified with a tag that facilitates purification
of the PCR products. The modification is preferably the same for
all primers, although different modifications can be used if it is
desired to separate the PCR products into different groups.
[0260] The tag can be any chemical moiety including but not limited
to a radioisotope, fluorescent reporter molecule, chemiluminescent
reporter molecule, antibody, antibody fragment, hapten, biotin,
derivative of biotin, photobiotin, iminobiotin, digoxigenin,
avidin, enzyme, acridinium, sugar, enzyme, apoenzyme, homopolymeric
oligonucleotide, hormone, ferromagnetic moiety, paramagnetic
moiety, diamagnetic moiety, phosphorescent moiety, luminescent
moiety, electrochemiluminescent moiety, chromatic moiety, moiety
having a detectable electron spin resonance, electrical
capacitance, dielectric constant or electrical conductivity, or
combinations thereof.
[0261] As one example, the 5' ends of the primers can be
biotinylated (Kandpal et al., Nucleic Acids Res. 18:1789-1795
(1990); Kaneoka et al., Biotechniques 10:30-34 (1991); Green et
al., Nucleic Acids Res. 18:6163-6164 (1990)). The biotin provides
an affinity tag that can be used to purify the copied DNA from the
genomic DNA or any other DNA molecules that are not of interest.
Biotinylated molecules can be purified using a streptavidin coated
matrix as shown in FIG. 1F, including but not limited to
Streptawell, transparent, High-Bind plates from Roche Molecular
Biochemicals (catalog number 1 645 692, as listed in Roche
Molecular Biochemicals, 2001 Biochemicals Catalog).
[0262] The PCR product of each locus of interest is placed into
separate wells of a Streptavidin coated plate. Alternatively, the
PCR products of the loci of interest can be pooled and placed into
a streptavidin coated matrix, including but not limited to the
Streptawell, transparent, High-Bind plates from Roche Molecular
Biochemicals (catalog number 1 645 692, as listed in Roche
Molecular Biochemicals, 2001 Biochemicals Catalog).
[0263] The amplified DNA can also be separated from the template
DNA using non-affinity methods known in the art, for example, by
polyacrylamide gel electrophoresis using standard protocols.
IV. Digestion of Amplified DNA
[0264] The amplified DNA can be digested with a restriction enzyme
that recognizes a sequence that had been provided on the first or
second primer using standard protocols known within the art (FIGS.
6A-6D). Restriction enzyme digestions are performed using standard
protocols well known within the art. The enzyme used depends on the
restriction recognition site generated with the first or second
primer. See "Primer Design" section, above, for details on
restriction recognition sites generated on primers.
[0265] Type IIS restriction enzymes are extremely useful in that
they cut approximately 10-20 base pairs outside of the recognition
site. Preferably, the Type IIS restriction enzymes used are those
that generate a 5' overhang and a recessed 3' end, including but
not limited to BceA I and BsmF I (see e.g. Table I). In a most
preferred embodiment, the second primer (either forward or reverse)
contains a restriction enzyme recognition sequence for BsmF I or
BceA I. The Type IIS restriction enzyme BsmF I recognizes the
nucleic acid sequence GGGAC, and cuts 14 nucleotides from the
recognition site on the antisense strand and 10 nucleotides from
the recognition site on the sense strand. Digestion with BsmF I
generates a 5' overhang of four (4) bases.
[0266] For example, if the second primer is designed so that after
amplification the restriction enzyme recognition site is 13 bases
from the locus of interest, then after digestion, the locus of
interest is the first base in the 5' overhang (reading 3' to 5'),
and the recessed 3' end is one base from the locus of interest. The
3' recessed end can be filled in with a nucleotide that is
complementary to the locus of interest. One base of the overhang
can be filled in using dideoxynucleotides. However, 1, 2, 3, or 4
bases of the overhang can be filled in using deoxynucleotides or a
mixture of dideoxynucleotides and deoxynucleotides.
[0267] The restriction enzyme BsmF I cuts DNA ten (10) nucleotides
from the recognition site on the sense strand and fourteen (14)
nucleotides from the recognition site on the antisense strand.
However, in a sequence dependent manner, the restriction enzyme
BsmF I also cuts eleven (11) nucleotides from the recognition site
on the sense strand and fifteen (15) nucleotides from the
recognition site on the antisense strand. Thus, two populations of
DNA molecules exist after digestion: DNA molecules cut at 10/14 and
DNA molecules cut at 11/15. If the recognition site for BsmF I is
13 bases from the locus of interest in the amplified product, then
DNA molecules cut at the 11/15 position will generate a 5' overhang
that contains the locus of interest in the second position of the
overhang (reading 3' to 5'). The 3' recessed end of the DNA
molecules can be filled in with labeled nucleotides. For example,
if labeled dideoxynucleotides are used, the 3' recessed end of the
molecules cut at 11/15 would be filled in with one base, which
corresponds to the base upstream from the locus of interest, and
the 3' recessed end of molecules cut at 10/14 would be filled in
with one base, which corresponds to the locus of interest. The DNA
molecules that have been cut at the 10/14 position and the DNA
molecules that have been cut at the 11/15 position can be separated
by size, and the incorporated nucleotides detected. This allows
detection of both the nucleotide before the locus of interest,
detection of the locus of interest, and potentially the three bases
after the locus of interest.
[0268] Alternatively, if the base upstream from the locus of
interest and the locus of interest are different nucleotides, then
the 3' recessed end of the molecules cut at 11/15 can be filled in
with deoxynucleotide that is complementary to the upstream base.
The remaining deoxynucleotide is washed away, and the locus of
interest site can be filled in with either labeled
deoxynucleotides, unlabeled deoxynucleotides, labeled
dideoxynucleotides, or unlabeled dideoxynucleotides. After the fill
in reaction, the nucleotide can be detected by any suitable method.
Thus, after the first fill in reaction with dNTP, the 3' recessed
end of the molecules cut at 10/14 and 11/15 is upstream from the
locus of interest. The 3' recessed end can now be filled in one
base, which corresponds to the locus of interest, two bases, three
bases or four bases.
[0269] The restriction enzyme BceA I recognizes the nucleic acid
sequence ACGGC and cuts 12 (twelve) nucleotides from the
recognition site on the sense strand and 14 (fourteen) nucleotides
from the recognition site on the antisense strand. If the distance
from the recognition site for BceA I on the second primer is
designed to be thirteen (13) bases from the locus of interest (see
FIGS. 4A-4D), digestion with BceA I will generate a 5' overhang of
two bases, which contains the locus of interest, and a recessed 3'
end that is upstream from the locus of interest. The locus of
interest is the first nucleotide in the 5' overhang (reading 3' to
5').
[0270] Alternative cutting is also seen with the restriction enzyme
BceA I, although at a much lower frequency than is seen with BsmF
I. The restriction enzyme BceA I can cut thirteen (13) nucleotides
from the recognition site on the sense strand and fifteen (15)
nucleotides from the recognition site on the antisense strand.
Thus, two populations of DNA molecules exist: DNA molecules cut at
12/14 and DNA molecules cut at 13/15. If the restriction enzyme
recognition site is 13 bases from the locus of interest in the
amplified product, DNA molecules cut at the 13/15 position yield a
5' overhang, which contains the locus of interest in the second
position of the overhang (reading 3' to 5'). Labeled
dideoxynucleotides can be used to fill in the 3' recessed end of
the DNA molecules. The DNA molecules cut at 13/15 will have the
base upstream from the locus of interest filled in, and the DNA
molecules cut at 12/14 will have the locus of interest site filled
in. The DNA molecules cut at 13/15 and those cut at 12/14 can be
separated by size, and the incorporated nucleotide detected. Thus,
the alternative cutting can be used to obtain additional sequence
information.
[0271] Alternatively, if the two bases in the 5' overhang are
different, the 3' recessed end of the DNA molecules, which were cut
at 13/15, can be filled in with the deoxynucleotide complementary
to the first base in the overhang, and excess deoxynucleotide
washed away. After filling in, the 3' recessed end of the DNA
molecules that were cut at 12/14 and the DNA molecules that were
cut at 13/15 are upstream from the locus of interest. The 3'
recessed ends can be filled with either labeled dideoxynucleotides,
unlabeled dideoxynucleotides, labeled deoxynucleotides, or
unlabeled deoxynucleotides.
[0272] If the primers provide different restriction sites for
certain of the loci of interest that were copied, all the necessary
restriction enzymes can be added together to digest the copied DNA
simultaneously. Alternatively, the different restriction digests
can be made in sequence, for example, using one restriction enzyme
at a time, so that only the product that is specific for that
restriction enzyme is digested.
[0273] Optimal restriction enzyme digestion conditions, including
but not limited to the concentration of enzyme, temperature, buffer
conditions, and the time of digestion can be optimized for each
restriction enzyme. For example, the alternative cutting seen with
the type IIS restriction enzyme BsmF I can be reduced, if desired,
by performing the restriction enzyme digestion at lower
temperatures including but not limited to 25-16.degree.,
16-12.degree. C., 12-8.degree. C., 8-4.degree. C., or 4-0.degree.
C.
V. Incorporation of Labeled Nucleotides
[0274] Digestion with the restriction enzyme that recognizes the
sequence on the second primer generates a recessed 3' end and a 5'
overhang, which contains the locus of interest (FIG. 1G). The
recessed 3' end can be filled in using the 5' overhang as a
template in the presence of unlabeled or labeled nucleotides or a
combination of both unlabeled and labeled nucleotides. The
nucleotides can be labeled with any type of chemical group or
moiety that allows for detection including but not limited to
radioactive molecules, fluorescent molecules, antibodies, antibody
fragments, haptens, carbohydrates, biotin, derivatives of biotin,
phosphorescent moieties, luminescent moieties,
electrochemiluminescent moieties, chromatic moieties, and moieties
having a detectable electron spin resonance, electrical
capacitance, dielectric constant or electrical conductivity. The
nucleotides can be labeled with one or more than one type of
chemical group or moiety. Each nucleotide can be labeled with the
same chemical group or moiety. Alternatively, each different
nucleotide can be labeled with a different chemical group or
moiety. The labeled nucleotides can be dNTPs, ddNTPs, or a mixture
of both dNTPs and ddNTPs. The unlabeled nucleotides can be dNTPs,
ddNTPs or a mixture of both dNTPs and ddNTPs.
[0275] Any combination of nucleotides can be used to incorporate
nucleotides including but not limited to unlabeled
deoxynucleotides, labeled deoxynucleotides, unlabeled
dideoxynucleotides, labeled dideoxynucleotides, a mixture of
labeled and unlabeled deoxynucleotides, a mixture of labeled and
unlabeled dideoxynucleotides, a mixture of labeled deoxynucleotides
and labeled dideoxynucleotides, a mixture of labeled
deoxynucleotides and unlabeled dideoxynucleotides, a mixture of
unlabeled deoxynucleotides and unlabeled dideoxynucleotides, a
mixture of unlabeled deoxynucleotides and labeled
dideoxynucleotides, dideoxynucleotide analogues, deoxynucleotide
analogues, a mixture of dideoxynucleotide analogues and
deoxynucleotide analogues, phosphorylated nucleoside analogues,
2'-deoxynucleotide-5'-triphosphate, and modified
2'-deoxynucleotide-5'-triphosphate.
[0276] For example, as shown in FIG. 1H, in the presence of a
polymerase, the 3' recessed end can be filled in with fluorescent
ddNTP using the 5' overhang as a template. The incorporated ddNTP
can be detected using any suitable method including but not limited
to fluorescence detection.
[0277] All four nucleotides can be labeled with different
fluorescent groups, which will allow one reaction to be performed
in the presence of all four labeled nucleotides. Alternatively,
four separate "fill in" reactions can be performed for each locus
of interest; each of the four reactions will contain a different
labeled nucleotide (e.g. ddATP*, ddTTP*, ddGTP*, or ddCTP*, where *
indicates a labeled nucleotide). Each nucleotide can be labeled
with different chemical groups or the same chemical groups. The
labeled nucleotides can be dideoxynucleotides or
deoxynucleotides.
[0278] In another embodiment, nucleotides can be labeled with
fluorescent dyes including but not limited to fluorescein, pyrene,
7-methoxycoumarin, Cascade Blue.TM., Alexa Flur 350, Alexa Flur
430, Alexa Flur 488, Alexa Flur 532, Alexa Flur 546, Alexa Flur
568, Alexa Flur 594, Alexa Flur 633, Alexa Flur 647, Alexa Flur
660, Alexa Flur 680, AMCA-X, dialkylaminocoumarin, Pacific Blue,
Marina Blue, BODIPY 493/503, BODIPY F1-X, DTAF, Oregon Green 500,
Dansyl-X, 6-FAM, Oregon Green 488, Oregon Green 514, Rhodamine
Green-X, Rhodol Green, Calcein, Eosin, ethidium bromide, NBD, TET,
2',4',5',7' tetrabromosulfonefluorescien, BODIPY R6G, BODIPY-F1
BR2, BODIPY 530/550, HEX, BODIPY 558/568, BODIPY-TMR-X., PyMPO,
BODIPY 564/570, TAMRA, BODIPY 576/589, Cy3, Rhodamine Red-x, BODIPY
581/591, carboxyXrhodamine, Texas Red-X, BODIPY-TR-X., Cy5,
SpectrumAqua, SpectrumGreen #1, SpectrumGreen #2, SpectrumOrange,
SpectrumRed, or naphthofluorescein.
[0279] In another embodiment, the "fill in" reaction can be
performed with fluorescently labeled dNTPs, wherein the nucleotides
are labeled with different fluorescent groups. The incorporated
nucleotides can be detected by any suitable method including but
not limited to Fluorescence Resonance Energy Transfer (FRET).
[0280] In another embodiment, a mixture of both labeled ddNTPs and
unlabeled dNTPs can be used for filling in the recessed 3' end of
the SNP or locus of interest. Preferably, the 5' overhang consists
of more than one base, including but not limited to 2, 3, 4, 5, 6
or more than 6 bases. For example, if the 5' overhang consists of
the sequence "XGAA," wherein X is the locus of interest, e.g. SNP,
then filling in with a mixture of labeled ddNTPs and unlabeled
dNTPs will produce several different DNA fragments. If a labeled
ddNTP is incorporated at position "X," the reaction will terminate
and a single labeled base will be incorporated. If however, an
unlabeled dNTP is incorporated, the polymerase continues to
incorporate other bases until a labeled ddNTP is incorporated. If
the first two nucleotides incorporated are dNTPs, and the third is
a ddNTP, the 3' recessed end will be extend by three bases. This
DNA fragment can be separated from the other DNA fragments that
were extended by 1, 2, or 4 bases by size. A mixture of labeled
ddNTPs and unlabeled dNTPs will allow all bases of the overhang to
be filled in, and provides additional sequence information about
the locus of interest, e.g. SNP (see FIGS. 7E and 9D).
[0281] After incorporation of the labeled nucleotide, the amplified
DNA can be digested with a restriction enzyme that recognizes the
sequence provided by the first primer. For example, in FIG. 1I, the
amplified DNA is digested with a restriction enzyme that binds to
region "a," which releases the DNA fragment containing the
incorporated nucleotide from the streptavidin matrix.
[0282] Alternatively, one primer of each primer pair for each locus
of interest can be attached to a solid support matrix including but
not limited to a well of a microliter plate. For example,
streptavidin-coated microtiter plates can be used for the
amplification reaction with a primer pair, wherein one primer is
biotinylated. First, biotinylated primers are bound to the
streptavidin-coated microtiter plates. Then, the plates are used as
the reaction vessel for PCR amplification of the loci of interest.
After the amplification reaction is complete, the excess primers,
salts, and template DNA can be removed by washing. The amplified
DNA remains attached to the microtiter plate. The amplified DNA can
be digested with a restriction enzyme that recognizes a sequence on
the second primer and generates a 5' overhang, which contains the
locus of interest. The digested fragments can be removed by
washing. After digestion, the SNP site or locus of interest is
exposed in the 5' overhang. The recessed 3' end is filled in with a
labeled nucleotide, including but not limited to, fluorescent ddNTP
in the presence of a polymerase. The labeled DNA can be released
into the supernatant in the microtiter plate by digesting with a
restriction enzyme that recognizes a sequence in the 5' region of
the first primer.
[0283] In another embodiment, one nucleotide can be used to
determine the sequence of multiple alleles of a gene. A nucleotide
that terminates the elongation reaction can be used to determine
the sequence of multiple alleles of a gene. At one allele, the
terminating nucleotide is complementary to the locus of interest in
the 5' overhang of said allele. The nucleotide is incorporated and
terminates the reaction. At a different allele, the terminating
nucleotide is not complementary to the locus of interest, which
allows a non-terminating nucleotide to be incorporated at the locus
of interest of the different allele. However, the terminating
nucleotide is complementary to a nucleotide downstream from the
locus of interest in the 5' overhang of said different allele. The
sequence of the alleles can be determined by analyzing the patterns
of incorporation of the terminating nucleotide. The terminating
nucleotide can be labeled or unlabeled.
[0284] In a another embodiment, the terminating nucleotide is a
nucleotide that terminates or hinders the elongation reaction
including but not limited to a dideoxynucleotide, a
dideoxynucleotide derivative, a dideoxynucleotide analog, a
dideoxynucleotide homolog, a dideoxynucleotide with a sulfur
chemical group, a deoxynucleotide, a deoxynucleotide derivative, a
deoxynucleotide homolog, a deoxynucleotide analog, a
deoxynucleotide with a sulfur chemical group, arabinoside
triphosphate, an arabinoside triphosphate analog, an arabinoside
triphosphate homolog, or an arabinoside derivative.
[0285] In another embodiment, a terminating nucleotide labeled with
one signal generating moiety tag, including but not limited to a
fluorescent dye, can be used to determine the sequence of the
alleles of a locus of interest. The use of a single nucleotide
labeled with one signal generating moiety tag eliminates any
difficulties that can arise when using different fluorescent
moieties. In addition, using one nucleotide labeled with one signal
generating moiety tag to determine the sequence of alleles of a
locus of interest reduces the number of reactions, and eliminates
pipetting errors.
[0286] For example, if the second primer contains the restriction
enzyme recognition site for BsmFI, digestion will generate a 5'
overhang of 4 bases. The second primer can be designed such that
the locus of interest is located in the first position of the
overhang. A representative overhang is depicted below, where R
represents the locus of interest:
TABLE-US-00008 5' CAC 3' GTG R T G G Overhang position 1 2 3 4
[0287] One nucleotide with one signal generating moiety tag can be
used to determine whether the variable site is homozygous or
heterozygous. For example, if the variable site is adenine (A) or
guanine (G), then either adenine or guanine can be used to
determine the sequence of the alleles of the locus of interest,
provided that there is an adenine or guanine in the overhang at
position 2, 3, or 4.
[0288] For example, if the nucleotide in position 2 of the overhang
is thymidine, which is complementary to adenine, then labeled
ddATP, unlabeled dCTP, dGTP, and dTTP can be used to determine the
sequence of the alleles of the locus of interest. The ddATP can be
labeled with any signal generating moiety including but not limited
to a fluorescent dye. If the template DNA is homozygous for
adenine, then labeled ddATP* will be incorporated at position 1
complementary to the overhang at the alleles, and no nucleotide
incorporation will be seen at position 2, 3 or 4 complementary to
the overhang.
TABLE-US-00009 Allele 1 5' CCC A* 3' GGG T T G G Overhang position
1 2 3 4 Allele 2 5' CCC A* 3' GGG T T G G Overhang position 1 2 3
4
[0289] One signal will be seen corresponding to incorporation of
labeled ddATP at position 1 complementary to the overhang, which
indicates that the individual is homozygous for adenine at this
position. This method of labeling eliminates any difficulties that
may arise from using different dyes that have different quantum
coefficients.
[0290] Homozygous Guanine:
[0291] If the template DNA is homozygous for guanine, then no ddATP
will be incorporated at position 1 complementary to the overhang,
but ddATP will be incorporated at the first available position,
which in this case is position 2 complementary to the overhang. For
example, if the second position in the overhang corresponds to a
thymidine, then:
TABLE-US-00010 Allele 1 5' CCC G A* 3' GGG C T G G Overhang
position 1 2 3 4 Allele 2 5' CCC G A* 3' GGG C T G G Overhang
position 1 2 3 4
[0292] One signal will be seen corresponding to incorporation of
ddATP at position 2 complementary to the overhang, which indicates
that the individual is homozygous for guanine. The molecules that
are filled in at position 2 complementary to the overhang will have
a different molecular weight than the molecules filled in at
position 1 complementary to the overhang.
[0293] Heterozygous Condition;
TABLE-US-00011 Allele 1 5' CCC A* 3' GGG T T G G Overhang position
1 2 3 4 Allele 2 5' CCC G A* 3' GGG C T G G Overhang position 1 2 3
4
[0294] Two signals will be seen; the first signal corresponds to
the ddATP filled in at position one complementary to the overhang
and the second signal corresponds to the ddATP filled in at
position 2 complementary to the overhang. The two signals can be
separated based on molecular weight; allele 1 and allele 2 will be
separated by a single base pair, which allows easy detection and
quantitation of the signals. Molecules filled in at position one
can be distinguished from molecules filled in at position two using
any method that discriminates based on molecular weight including
but not limited to gel electrophoresis, capillary gel
electrophoresis, DNA sequencing, and mass spectrometry. It is not
necessary that the nucleotide be labeled with a chemical moiety;
the DNA molecules corresponding to the different alleles can be
separated based on molecular weight.
[0295] If position 2 of the overhang is not complementary to
adenine, it is possible that positions 3 or 4 may be complementary
to adenine. For example, position 3 of the overhang may be
complementary to the nucleotide adenine, in which case labeled
ddATP may be used to determine the sequence of both alleles.
[0296] Homozygous for Adenine:
TABLE-US-00012 Allele 1 5' CCC A* 3' GGG T G T G Overhang position
1 2 3 4 Allele 2 5' CCC A* 3' GGG T G T G Overhang position 1 2 3
4
[0297] Homozygous for Guanine:
TABLE-US-00013 Allele 1 5' CCC G C A* 3' GGG C G T G Overhang
position 1 2 3 4 Allele 2 5' CCC G C A* 3' GGG C G T G Overhang
position 1 2 3 4
[0298] Heterozygous:
TABLE-US-00014 Allele 1 5' CCC A* 3' GGG T G T G Overhang position
1 2 3 4 Allele 2 5' CCC G C A* 3' GGG C G T G Overhang position 1 2
3 4
[0299] Two signals will be seen; the first signal corresponds to
the ddATP filled in at position 1 complementary to the overhang and
the second signal corresponds to the ddATP filled in at position 3
complementary to the overhang. The two signals can be separated
based on molecular weight; allele 1 and allele 2 will be separated
by two bases, which can be detected using any method that
discriminates based on molecular weight.
[0300] Alternatively, if positions 2 and 3 are not complementary to
adenine (i.e positions 2 and 3 of the overhang correspond to
guanine, cytosine, or adenine) but position 4 is complementary to
adenine, labeled ddATP can be used to determine the sequence of
both alleles.
[0301] Homozygous for Adenine:
TABLE-US-00015 Allele 1 5' CCC A* 3' GGG T G G T Overhang position
1 2 3 4 Allele 2 5' CCC A* 3' GGG T G G T Overhang position 1 2 3
4
[0302] One signal will be seen that corresponds to the molecular
weight of molecules filled in with ddATP at position one
complementary to the overhang, which indicates that the individual
is homozygous for adenine at the variable site.
[0303] Homozygous for Guanine:
TABLE-US-00016 Allele 1 5' CCC G C C A* 3' GGG C G G T Overhang
position 1 2 3 4 Allele 2 5' CCC G C C A* 3' GGG C G G T Overhang
position 1 2 3 4
[0304] One signal will be seen that corresponds to the molecular
weight of molecules filled in at position 4 complementary to the
overhang, which indicates that the individual is homozygous for
guanine.
[0305] Heterozygous:
TABLE-US-00017 Allele 1 5' CCC A* 3' GGG T G G T Overhang position
1 2 3 4 Allele 2 5' CCC G C C A* 3' GGG C G G T Overhang position 1
2 3 4
[0306] Two signals will be seen; the first signal corresponds to
the ddATP filled in at position one complementary to the overhang
and the second signal corresponds to the ddATP filled in at
position 4 complementary to the overhang. The two signals can be
separated based on molecular weight; allele 1 and allele 2 will be
separated by three bases, which allows detection and quantitation
of the signals. The molecules filled in at position 1 and those
filled in at position 4 can be distinguished based on molecular
weight.
[0307] As discussed above, if the variable site contains either
adenine or guanine, either labeled adenine or labeled guanine can
be used to determine the sequence of both alleles. If positions 2,
3, or 4 of the overhang are not complementary to adenine but one of
the positions is complementary to a guanine, then labeled ddGTP can
be used to determine whether the template DNA is homozygous or
heterozygous for adenine or guanine. For example, if position 3 in
the overhang corresponds to a cytosine then the following signals
will be expected if the template DNA is homozygous for guanine,
homozygous for adenine, or heterozygous:
[0308] Homozygous for Guanine:
TABLE-US-00018 Allele 1 5' CCC G* 3' GGG C T C T Overhang position
1 2 3 4 Allele 2 5' CCC G* 3' GGG C T C T Overhang position 1 2 3
4
[0309] One signal will be seen that corresponds to the molecular
weight of molecules filled in with ddGTP at position one
complementary to the overhang, which indicates that the individual
is homozygous for guanine.
[0310] Homozygous for Adenine:
TABLE-US-00019 Allele 1 5' CCC A A G* 3' GGG T T C T Overhang
position 1 2 3 4 Allele 2 5' CCC A A G* 3' GGG T T C T Overhang
position 1 2 3 4
[0311] One signal will be seen that corresponds to the molecular
weight of molecules filled in at position 3 complementary to the
overhang, which indicates that the individual is homozygous for
adenine at the variable site.
[0312] Heterozygous:
TABLE-US-00020 Allele 1 5' CCC G* 3' GGG C T C T Overhang position
1 2 3 4 Allele 2 5' CCC A A G* 3' GGG T T C T Overhang position 1 2
3 4
[0313] Two signals will be seen; the first signal corresponds to
the ddGTP filled in at position one complementary to the overhang
and the second signal corresponds to the ddGTP filled in at
position 3 complementary to the overhang. The two signals can be
separated based on molecular weight; allele 1 and allele 2 will be
separated by two bases, which allows easy detection and
quantitation of the signals.
[0314] In another embodiment, the nucleotide labeled with a single
chemical moiety, which is used to determine the sequence of alleles
of interest, can be analyzed by a variety of methods including but
not limited to fluorescence detection, DNA sequencing gel,
capillary electrophoresis on an automated DNA sequencing machine,
microchannel electrophoresis, and other methods of sequencing, mass
spectrometry, time of flight mass spectrometry, quadrupole mass
spectrometry, magnetic sector mass spectrometry, electric sector
mass spectrometry infrared spectrometry, ultraviolet spectrometry,
palentiostatic amperometry or by DNA hybridization techniques
including Southern Blots, Slot Blots, Dot Blots, and DNA
microarrays, wherein DNA fragments would be useful as both "probes"
and "targets," ELISA, fluorimetry, Fluorescence Resonance Energy
Transfer (FRET), SNP-IT, GeneChips, HuSNP, BeadArray, TaqMan assay,
Invader assay, MassExtend, or MassCleave.TM. (hMC) method.
[0315] Some type IIS restriction enzymes also display alternative
cutting as discussed above. For example, BsmFI will cut at 10/14
and 11/15 from the recognition site. However, the cutting patterns
are not mutually exclusive; if the 11/15 cutting pattern is seen at
a particular sequence, 10/14 cutting is also seen. If the
restriction enzyme BsmF I cuts at 10/14 from the recognition site,
the 5' overhang will be X.sub.1X.sub.2X.sub.3X.sub.4. If BsmF I
cuts 11/15 from the recognition site, the 5' overhang will be
X.sub.0X.sub.1X.sub.2X.sub.3. If position X.sub.0 of the overhang
is complementary to the labeled nucleotide, the labeled nucleotide
will be incorporated at position X.sub.0 and provides an additional
level of quality assurance. It provides additional sequence
information.
[0316] For example, if the variable site is adenine or guanine, and
position 3 in the overhang is complementary to adenine, labeled
ddATP can be used to determine the genotype at the variable site.
If position 0 of the 11/15 overhang contains the nucleotide
complementary to adenine, ddATP will be filled in and an additional
signal will be seen.
[0317] Heterozygous:
TABLE-US-00021 10/14 Allele 1 5' CCA A* 3' GGT T G T G Overhang
position 1 2 3 4 10/14 Allele 2 5' CCA G C A* 3' GGT C G T G
Overhang position 1 2 3 4 11/15 Allele 1 5' CC A* 3' GG T T G T
Overhang position 0 1 2 3 11/15 Allele 2 5' CC A* 3' GG T C G T
Overhang position 0 1 2 3
[0318] Three signals are seen; one corresponding to the ddATP
incorporated at position 0 complementary to the overhang, one
corresponding to the ddATP incorporated at position 1 complementary
to the overhang, and one corresponding to the ddATP incorporated at
position 3 complementary to the overhang. The molecules filled in
at position 0, 1, and 3 complementary to the overhang differ in
molecular weight and can be separated using any technique that
discriminates based on molecular weight including but not limited
to gel electrophoresis, and mass spectrometry.
[0319] For quantitating the ratio of one allele to another allele
or when determining the relative amount of a mutant DNA sequence in
the presence of wild type DNA sequence, an accurate and highly
sensitive method of detection must be used. The alternate cutting
displayed by type IIS restriction enzymes may increase the
difficulty of determining ratios of one allele to another allele
because the restriction enzyme may not display the alternate
cutting (11/15) pattern on the two alleles equally. For example,
allele 1 may be cut at 10/14 80% of the time, and 11/15 20% of the
time. However, because the two alleles may differ in sequence,
allele 2 may be cut at 10/14 90% of the time, and 11/15 20% of the
time.
[0320] For purposes of quantitation, the alternate cutting problem
can be eliminated when the nucleotide at position 0 of the overhang
is not complementary to the labeled nucleotide. For example, if the
variable site corresponds to adenine or guanine, and position 3 of
the overhang is complementary to adenine (i.e, a thymidine is
located at position 3 of the overhang), labeled ddATP can be used
to determine the genotype of the variable site. If position 0 of
the overhang generated by the 11/15 cutting properties is not
complementary to adenine, (i.e, position 0 of the overhang
corresponds to guanine, cytosine, or adenine) no additional signal
will be seen from the fragments that were cut 11/15 from the
recognition site. Position 0 complementary to the overhang can be
filled in with unlabeled nucleotide, eliminating any complexity
seen from the alternate cutting pattern of restriction enzymes.
This method provides a highly accurate method for quantitating the
ratio of a variable site including but not limited to a mutation,
or a single nucleotide polymorphism.
[0321] For instance, if SNP X can be adenine or guanine, this
method of labeling allows quantitation of the alleles that
correspond to adenine and the alleles that correspond to guanine,
without determining if the restriction enzyme displays any
differences between the alleles with regard to alternate cutting
patterns.
[0322] Heterozygous:
TABLE-US-00022 10/14 Allele 1 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4 10/14 Allele 2 5' CCG G C A* 3' GGC C G T G
Overhang position 1 2 3 4
[0323] The overhang generated by the alternate cutting properties
of BsmF I is depicted below:
TABLE-US-00023 11/15 Allele 1 5' CC 3' GG C T G T Overhang position
0 1 2 3 11/15 Allele 2 5' CC 3' GG C C G T Overhang position 0 1 2
3
[0324] After filling in with labeled ddATP and unlabeled dGTP,
dCTP, dTTP, the following molecules would be generated:
TABLE-US-00024 11/15 Allele 1 5' CC G A* 3' GG C T G T Overhang
position 0 1 2 3 11/15 Allele 2 5' CC G G C A* 3' GG C C G T
Overhang position 0 1 2 3
[0325] Two signals are seen; one corresponding to the molecules
filled in with ddATP at position one complementary to the overhang
and one corresponding to the molecules filled in with ddATP at
position 3 complementary to the overhang. Position 0 of the 11/15
overhang is filled in with unlabeled nucleotide, which eliminates
any difficulty in quantitating a ratio for the nucleotide at the
variable site on allele 1 and the nucleotide at the variable site
on allele 2.
[0326] Any nucleotide can be used including adenine, adenine
derivatives, adenine homologues, guanine, guanine derivatives,
guanine homologues, cytosine, cytosine derivatives, cytosine
homologues, thymidine, thymidine derivatives, or thymidine
homologues, or any combinations of adenine, adenine derivatives,
adenine homologues, guanine, guanine derivatives, guanine
homologues, cytosine, cytosine derivatives, cytosine homologues,
thymidine, thymidine derivatives, or thymidine homologues.
[0327] The nucleotide can be labeled with any chemical group or
moiety, including but not limited to radioactive molecules,
fluorescent molecules, antibodies, antibody fragments, haptens,
carbohydrates, biotin, derivatives of biotin, phosphorescent
moieties, luminescent moieties, electrochemiluminescent moieties,
chromatic moieties, and moieties having a detectable electron spin
resonance, electrical capacitance, dielectric constant or
electrical conductivity. The nucleotide can be labeled with one or
more than one type of chemical group or moiety.
[0328] In another embodiment, labeled and unlabeled nucleotides can
be used. Any combination of deoxynucleotides and dideoxynucleotides
can be used including but not limited to labeled dideoxynucleotides
and labeled deoxynucleotides; labeled dideoxynucleotides and
unlabeled deoxynucleotides; unlabeled dideoxynucleotides and
unlabeled deoxynucleotides; and unlabeled dideoxynucleotides and
labeled deoxynucleotides.
[0329] In another embodiment, nucleotides labeled with a chemical
moiety can be used in the PCR reaction. Unlabeled nucleotides then
are used to fill-in the 5' overhangs generated after digestion with
the restriction enzyme. An unlabeled terminating nucleotide can be
used to in the presence of unlabeled nucleotides to determine the
sequence of the alleles of a locus of interest.
[0330] For example, if labeled dTTP was used in the PCR reaction,
the following 5' overhang would be generated after digestion with
BsmF I:
TABLE-US-00025 10/14 Allele 1 5' CT*G A 3' GA C T G T G Overhang
position 1 2 3 4 10/14 Allele 2 5' CT*G G C A 3' GA C C G T G
Overhang position 1 2 3 4
[0331] Unlabeled ddATP, unlabeled dCTP, unlabeled dGTP, and
unlabeled dTTP can be used to fill-in the 5' overhang. Two signals
will be generated; one signal corresponds to the DNA molecules
filled in with unlabeled ddATP at position 1 complementary to the
overhang and the second signal corresponds to DNA molecules filled
in with unlabeled ddATP at position 3 complementary to the
overhang. The DNA molecules can be separated based on molecular
weight and can be detected by the fluorescence of the dTTP, which
was incorporated during the PCR reaction.
[0332] The labeled DNA loci of interest sites can be analyzed by a
variety of methods including but not limited to fluorescence
detection, DNA sequencing gel, capillary electrophoresis on an
automated DNA sequencing machine, microchannel electrophoresis, and
other methods of sequencing, mass spectrometry, time of flight mass
spectrometry, quadrupole mass spectrometry, magnetic sector mass
spectrometry, electric sector mass spectrometry infrared
spectrometry, ultraviolet spectrometry, palentiostatic amperometry
or by DNA hybridization techniques including Southern Blots, Slot
Blots, Dot Blots, and DNA microarrays, wherein DNA fragments would
be useful as both "probes" and "targets," ELISA, fluorimetry,
Fluorescence Resonance Energy Transfer (FRET), SNP-IT, GeneChips,
HuSNP, BeadArray, TaqMan assay, Invader assay, MassExtend, or
MassCleave.TM. (hMC) method.
[0333] This method of labeling is extremely sensitive and allows
the detection of alleles of a locus of interest that are in various
ratios including but not limited to 1:1, 1:2, 1:3, 1:4, 1:5,
1:6-1:10, 1:11-1:20, 1:21-1:30, 1:31-1:40, 1:41-1:50, 1:51-1:60,
1:61-1:70, 1:71-1:80, 1:81-1:90, 1:91:1:100, 1:101-1:200, 1:250,
1:251-1:300, 1:301-1:400, 1:401-1:500, 1:501-1:600, 1:601-1:700,
1:701-1:800, 1:801-1:900, 1:901-1:1000, 1:1001-1:2000,
1:2001-1:3000, 1:3001-1:4000, 1:4001-1:5000, 1:5001-1:6000,
1:6001-1:7000, 1:7001-1:8000, 1:8001-1:9000, 1:9001-1:10,000;
1:10,001-1:20,000, 1:20,001:1:30,000, 1:30,001-1:40,000,
1:40,001-1:50,000, and greater than 1:50,000.
[0334] For example, this method of labeling allows one nucleotide
labeled with one signal generating moiety to be used to determine
the sequence of alleles at a SNP locus, or detect a mutant allele
amongst a population of normal alleles, or detect an allele
encoding antibiotic resistance from a bacterial cell amongst
alleles from antibiotic sensitive bacteria, or detect an allele
from a drug resistant virus amongst alleles from drug-sensitive
virus, or detect an allele from a non-pathogenic bacterial strain
amongst alleles from a pathogenic bacterial strain.
[0335] As shown above, a single nucleotide can be used to determine
the sequence of the alleles at a particular locus of interest. This
method is especially useful for determining if an individual is
homozygous or heterozygous for a particular mutation or to
determine the sequence of the alleles at a particular SNP site.
This method of labeling eliminates any errors caused by the quantum
coefficients of various dyes. It also allows the reaction to
proceed in a single reaction vessel including but not limited to a
well of a microtiter plate, or a single eppendorf tube.
[0336] This method of labeling is especially useful for the
detection of multiple genetic signals in the same sample. For
example, this method is useful for the detection of fetal DNA in
the blood, serum, or plasma of a pregnant female, which contains
both maternal DNA and fetal DNA. The maternal DNA and fetal DNA may
be present in the blood, serum or plasma at ratios such as 97:3;
however, the above-described method can be used to detect the fetal
DNA. This method of labeling can be used to detect two, three, four
or more than four different genetic signals in the sample
population
[0337] This method of labeling is especially useful for the
detection of a mutant allele that is among a large population of
wild type alleles. Furthermore, this method of labeling allows the
detection of a single mutant cell in a large population of wild
type cells. For example, this method of labeling can be used to
detect a single cancerous cell among a large population of normal
cells. Typically, cancerous cells have mutations in the DNA
sequence. The mutant DNA sequence can be identified even if there
is a large background of wild type DNA sequence. This method of
labeling can be used to screen, detect, or diagnosis any type of
cancer including but not limited to colon, renal, breast, bladder,
liver, kidney, brain, lung, prostate, and cancers of the blood
including leukemia.
[0338] This labeling method can also be used to detect pathogenic
organisms, including but not limited to bacteria, fungi, viruses,
protozoa, and mycobacteria. It can also be used to discriminate
between pathogenic strains of microorganism and non-pathogenic
strains of microorganisms including but not limited to bacteria,
fungi, viruses, protozoa, and mycobacteria.
[0339] For example, there are several strains of Escherichia coii
(E. coli), and most are non-pathogenic. However, several strains,
such as E. coli O 157 are pathogenic. There are genetic differences
between non-pathogenic E. coli strains and pathogenic E. coli. The
above described method of labeling can be used to detect pathogenic
microorganisms in a large population of non-pathogenic organisms,
which are sometimes associated with the normal flora of an
individual.
VI. Analysis of the Locus of Interest
[0340] The loci of interest can be analyzed by a variety of methods
including but not limited to fluorescence detection, DNA sequencing
gel, capillary electrophoresis on an automated DNA sequencing
machine, (e.g. the ABI Prism 3100 Genetic Analyzer or the ABI Prism
3700 Genetic Analyzer), microchannel electrophoresis, and other
methods of sequencing, Sanger dideoxy sequencing, mass
spectrometry, time of flight mass spectrometry, quadrupole mass
spectrometry, magnetic sector mass spectrometry, electric sector
mass spectrometry infrared spectrometry, ultraviolet spectrometry,
palentiostatic amperometry or by DNA hybridization techniques
including Southern Blot, Slot Blot, Dot Blot, and DNA microarray,
wherein DNA fragments would be useful as both "probes" and
"targets," ELISA, fluorimetry, fluorescence polarization,
Fluorescence Resonance Energy Transfer (FRET), SNP-IT, GeneChips,
HuSNP, BeadArray, TaqMan assay, Invader assay, MassExtend, or
MassCleave.TM. (hMC) method.
[0341] The loci of interest can be analyzed using gel
electrophoresis followed by fluorescence detection of the
incorporated nucleotide. Another method to analyze or read the loci
of interest is to use a fluorescent plate reader or fluorimeter
directly on the 96-well streptavidin coated plates. The plate can
be placed onto a fluorescent plate reader or scanner such as the
Pharmacia 9200 Typhoon to read each locus of interest.
[0342] Alternatively, the PCR products of the loci of interest can
be pooled and after "filling in" (FIG. 10), the products can be
separated by size, using any method appropriate for the same, and
then analyzed using a variety of techniques including but not
limited to fluorescence detection, DNA sequencing gel, capillary
electrophoresis on an automated DNA sequencing machine,
microchannel electrophoresis, other methods of sequencing, Sanger
dideoxy sequencing, DNA hybridization techniques including Southern
Blot, Slot Blot, Dot Blot, and DNA microarray, mass spectrometry,
time of flight mass spectrometry, quadrupole mass spectrometry,
magnetic sector mass spectrometry, electric sector mass
spectrometry infrared spectrometry, ultraviolet spectrometry,
palentiostatic amperometry. For example, polyacrylamide gel
electrophoresis can be used to separate DNA by size and the gel can
be scanned to determine the color of fluorescence in each band
(using e.g., ABI 377 DNA sequencing machine or a Pharmacia Typhoon
9200).
[0343] In another embodiment, the sequence of the locus of interest
can be determined by detecting the incorporation of a nucleotide
that is 3' to the locus of interest, wherein said nucleotide is a
different nucleotide from the possible nucleotides at the locus of
interest. This embodiment is especially useful for the sequencing
and detection of SNPs. The efficiency and rate at which DNA
polymerases incorporate nucleotides varies for each nucleotide.
[0344] According to the data from the Human Genome Project, 99% of
all SNPs are binary. The sequence of the human genome can be used
to determine a nucleotide that is 3' to the SNP of interest. When a
nucleotide that is 3' to the SNP site differs from the possible
nucleotides at the SNP site, a nucleotide that is one or more than
one base 3' to the SNP can be used to determine the sequence of the
SNP site.
[0345] For example, suppose the sequence of SNP X on chromosome 13
is to be determined. The sequence of the human genome indicates
that SNP X can either be adenosine or guanine and that a nucleotide
3' to the locus of interest is a thymidine. A primer that contains
a restriction enzyme recognition site for BsmF I, which is designed
to be 13 bases from the locus of interest after amplification, is
used to amplify a DNA fragment containing SNP X. Digestion with the
restriction enzyme BsmF I generates a 5' overhang that contains the
locus of interest, which can either be adenosine or guanine. The
digestion products can be split into two "fill in" reactions: one
contains dTTP, and the other reaction contains dCTP. If the locus
of interest is homozygous for guanine, only the DNA molecules that
were mixed with dCTP will be filled in. If the locus of interest is
homozygous for adenosine, only the DNA molecules that were mixed
with dTTP will be filled in. If the locus of interest is
heterozygous, the DNA molecules that were mixed with dCTP will be
filled in as well as the DNA molecules that were mixed with dTTP.
After washing to remove the excess dNTP, the samples are filled in
with labeled ddATP, which is complimentary to the nucleotide
(thymidine) that is 3' to the locus of interest. The DNA molecules
that were filled in by the previous reaction will be filled in with
labeled ddATP. If the individual is homozygous for adenosine, the
DNA molecules that were mixed with dTTP subsequently will be filled
in with the labeled ddATP. However, the DNA molecules that were
mixed with dCTP, would not have incorporated that nucleotide, and
therefore, could not incorporate the ddATP. Detection of labeled
ddATP only in the molecules that were mixed with dTTP indicates
that the nucleotide at SNP X on chromosome 13 is adenosine.
[0346] In another embodiment, large scale screening for the
presence or absence of single nucleotide polymorphisms or mutations
can be performed. One to tens to hundreds to thousands of loci of
interest on a single chromosome or on multiple chromosomes can be
amplified with primers as described above in the "Primer Design"
section. The primers can be designed so that each amplified loci of
interest is of a different size (FIG. 2). The multiple loci of
interest can be of a DNA sample from one individual representing
multiple loci of interest on a single chromosome, multiple
chromosomes, multiple genes, a single gene, or any combination
thereof.
[0347] When human data is being analyzed, the known sequence can be
a specific sequence that has been determined from one individual
(including e.g. the individual whose DNA is currently being
analyzed), or it can be a consensus sequence such as that published
as part of the human genome.
Ratio of Alleles at Heterozygous Locus of Interest
[0348] In one embodiment, the ratio of alleles at a heterozygous
locus of interest can be calculated. The intensity of a nucleotide
at the loci of interest can be quantified using any number of
computer programs including but not limited to GeneScan and
ImageQuant. For example, for a heterozygous SNP, there are two
nucleotides, and each should be present in a 1:1 ratio. In a
preferred embodiment, the ratio of multiple heterozygous SNPs can
be calculated.
[0349] In one embodiment, the ratio for a variable nucleotide at
alleles at a heterozygous locus of interest can be calculated. The
intensity of each variable nucleotide present at the loci of
interest can be quantified using any number of computer programs
including but not limited to GeneScan and ImageQuant. For example,
for a heterozygous SNP, there will be two nucleotides present, and
each may be present in a 1:1 ratio. In a preferred embodiment, the
ratio of multiple heterozygous SNPs can be calculated.
[0350] In another embodiment, the ratio of alleles at a
heterozygous locus of interest on a chromosome is summed and
compared to the ratio of alleles at a heterozygous locus of
interest on a different chromosome. In a preferred embodiment, the
ratio of alleles at multiple heterozygous loci of interest on a
chromosome is summed and compared to the ratio of alleles at
multiple heterozygous loci of interest on a different chromosome.
The ratio obtained from SNP 1, SNP 2, SNP 3, SNP 4, etc on
chromosome 1 can be summed. This ratio can then be compared to the
ratio obtained from SNP A, SNP B, SNP C, SNP D, etc.
[0351] For example, 100 SNPs can be analyzed on chromosome 1. Of
these 100 SNPs, assume 50 are heterozygous. The ratio of the
alleles at heterozygous SNPs on chromosome 1 can be summed, and
should give a ratio of approximately 50:50. Likewise, of 100 SNPs
analyzed on chromosome 21, assume 50 are heterozygous. The ratio of
alleles at heterozygous SNPs on chromosome 21 is summed. With a
normal number of chromosomes, the ratio should be approximately
50:50, and thus there should be no difference between the ratio
obtained from chromosome 1 and 21. However, if there is an
additional copy of chromosome 21, an additional allele will be
provided, and the ratio should be approximately 66:33. Thus, the
ratio for nucleotides at heterozygous SNPs can be used to detect
the presence or absence of chromosomal abnormalities. Any
chromosomal abnormality can be detected including aneuploidy,
polyploidy, inversion, a trisomy, a monosomy, duplication,
deletion, deletion of a part of a chromosome, addition, addition of
a part of chromosome, insertion, a fragment of a chromosome, a
region of a chromosome, chromosomal rearrangement, and
translocation. The method is especially useful for the detection of
trisomy 13, trisomy 18, trisomy 21, XXY, and XYY.
[0352] The present invention provides a method to quantitate a
ratio for the alleles at a heterozygous locus of interest. The loci
of interest include but are not limited to single nucleotide
polymorphisms, mutations. There is no need to amplify the entire
sequence of a gene or to quantitate the amount of a particular gene
product. The present invention does not rely on quantitative
PCR.
Detection of Fetal Chromosomal Abnormalities
[0353] As discussed above in the section entitled "DNA template,"
the template DNA can be obtained from a sample of a pregnant
female, wherein the template DNA comprises maternal template DNA
and fetal template DNA. In one embodiment, the template DNA is
obtained from the blood of a pregnant female. In a preferred
embodiment, the template DNA is obtained from the plasma or serum
from the blood of a pregnant female.
[0354] In one embodiment, the template DNA from the sample from the
pregnant female comprises both maternal template DNA and fetal
template DNA. In another embodiment, maternal template DNA is
obtained from any nucleic acid containing source including but not
limited to cell, tissue, blood, serum, plasma, saliva, urine,
tears, vaginal secretion, lymph fluid, cerebrospinal fluid, mucosa
secretion, peritoneal fluid, ascitic fluid, fecal matter, or body
exudates, and sequenced to identify homozygous or heterozygous loci
of interest, which are the loci of interest analyzed on the
template DNA obtained from the sample from the pregnant female.
[0355] In a preferred embodiment, the sequence of the alleles of
multiple loci of interest on maternal template DNA is determined to
identify homozygous loci of interest. In another embodiment, the
sequence of the alleles of multiple loci of interest on maternal
template DNA is determined to identify heterozygous loci of
interest. The sequence of the alleles of multiple loci of interest
on maternal template DNA can be determined in a single reaction or
in multiple reactions.
[0356] For example, if 100 maternal loci of interest on chromosome
21 and 100 maternal loci of interest on chromosome 1 are analyzed,
one would predict approximately 50 loci of interest on each
chromosome to be homozygous and 50 to be heterozygous. The 50
homozygous loci of interest, or the 50 heterozygous loci of
interest or the 50 homozygous and 50 heterozygous loci of interest,
or any combination of the homozygous and heterozygous loci or
interest on each chromosome can be analyzed using the template DNA
from the sample from the pregnant female.
[0357] The locus of interest on the template DNA from the sample of
the pregnant female is analyzed using the amplification, isolation,
digestion, fill in, and detection methods described above. The same
primers used to analyze the locus of interest on the maternal
template DNA are used to screen the template DNA from the sample
from the pregnant female. Any number of loci of interest can be
analyzed on the template DNA from the sample from the pregnant
female. For example, 1, 1-5, 5-10, 10-20, 20-30, 30-40, 40-50,
50-60, 60-70, 70-80, 80-90, 90-100, 100-150, 150-200, 200-250,
250-300, 300-500, 500-1000, 1000-2000, 2000-3000, 3000-4000 or more
than 4000 homozygous maternal loci of interest can be analyzed in
the template DNA from the sample from the pregnant female. In a
preferred embodiment, multiple loci of interest on multiple
chromosomes are analyzed.
[0358] From the population of homozygous maternal loci of interest,
there will be both heterozygous and homozygous loci of interest
from the template DNA from the sample from the pregnant female; the
heterozygous loci of interest can be further analyzed. At
heterozygous loci of interest, the ratio of alleles can be used to
determine the number of chromosomes that are present.
[0359] The percentage of fetal DNA present in the sample from the
pregnant female can be calculated by determining the ratio of
alleles at a heterozygous locus of interest on a chromosome that is
not typically associated with a chromosomal abnormality. In a
preferred embodiment, the ratio of alleles at multiple heterozygous
loci of interest on a chromosome can be used to determine the
percentage of fetal DNA. For example, chromosome 1, which is the
largest chromosome in the human genome, can be used to determine
the percentage of fetal DNA.
[0360] For example, suppose SNP X is homozygous at the maternal
template DNA (A/A). At SNP X, the template DNA from the sample from
the pregnant female, which can contain both fetal DNA and maternal
DNA, is heterozygous (A/G). The nucleotide guanine represents the
fetal DNA because at SNP X the mother is homozygous, and thus the
guanine is attributed to the fetal DNA. The guanine at SNP X can be
used to calculate the percentage of fetal DNA in the sample.
[0361] Alternatively, multiple loci of interest on two or more
chromosomes can be examined to determine the percentage of fetal
DNA. For example, multiple loci of interest can be examined on
chromosomes 13, and 18 to determine the percentage of fetal DNA
because organisms with chromosomal abnormalities at chromosome 13
and 18 are not viable.
[0362] Alternatively, for a male fetus, a marker on the Y
chromosome can be used to determine the amount of fetal DNA present
in the sample. A panel of serial dilutions can be made using the
template DNA isolated from the sample from the pregnant female, and
quantitative PCR analysis performed. Two PCR reactions can be
performed: one PCR reaction to amplify a marker on the Y
chromosome, for example SRY, and the other reaction to amplify a
region on any of the autosomal chromosomes. The amount of fetal DNA
can be calculated using the following formula:
Percent Fetal DNA: (last dilution Y chromosome detected/last
dilution autosomal chromosome detected)*2*100.
[0363] If at SNP A, the mother is homozygous A/A, and the fetus is
heterozygous A/G, then the ratio of A:G can be used to detect
chromosomal abnormalities. If the fetal DNA is fifty percent (50%)
of the DNA in the maternal blood, then at SNP A where the maternal
nucleotide is an adenine and the other nucleotide is a guanine, one
would expect the ratio of adenine (two adenines from the maternal
template DNA and one from the fetal template DNA) to guanine (from
the fetal template DNA) to be 25:75 or 0.33. However, if the fetus
has a trisomy of this particular chromosome, and the additional
chromosome is contributed by the mother, and thus an additional
adenine nucleotide is present, then one would expect the ratio of
0.25 (50 (G)/(2*50 maternal A+2*50 fetal A). Thus, there is a
difference of 8% between the ratio obtained from a chromosome
present in two copies, and a chromosome present in a trisomy
condition. On the other hand, if the additional chromosome is
contributed by the father, and thus, an additional guanine is
present, then one would expect the ratio of 0.66 (2*50 for G fetal
allele/(2*50 maternal A allele+50 for fetal A allele).
[0364] However, if the fetal DNA is 40% of the DNA in the maternal
blood, the expected ratio without a trisomy is 0.25 (40 for fetal G
allele/2*60 for maternal A allele+1*60 for fetal A allele). If the
fetus has a trisomy, and the additional chromosome is provided by
the mother, the expected ratio would be 0.20 (40 for fetal G
allele/(2*60 for maternal A allele+2*40 for fetal A allele). A 5%
difference between the ratios obtained from a chromosome present in
two copies and a chromosome present in the Trisomy condition is
detected.
[0365] In another embodiment, multiple loci of interest on multiple
chromosomes can be examined. The ratios for the alleles at each
heterozygous locus of interest on a chromosome can be summed and
compared to the ratios for the alleles at each locus of interest on
a different chromosome. The chromosomes that are compared can be of
human origin, and include but are not limited to chromosomes 1, 2,
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
21, 22, X, and Y. The ratio obtained from multiple chromosomes can
be compared to the ratio obtained for a single chromosome or from
multiple chromosomes.
[0366] In one embodiment, one of the chromosomes used in the
comparison can be chromosome 13, 15, 16; 18, 21, 22, X or Y. In a
preferred embodiment, the ratios on chromosomes 13, 18, and 21 are
compared.
[0367] For example, assuming 40% fetal DNA in the sample from the
pregnant female, the ratio of the alleles at a heterozygous locus
of interest on chromosome 1 will be 0.25 (40 for fetal G
allele/(2*60 for maternal A allele+40 for fetal A allele).
Likewise, the ratio of alleles at a heterozygous locus of interest
on chromosome 21 will be present in a ratio of 0.25. However, in a
fetus with trisomy 21 where the additional chromosome is
contributed by the mother, the nucleotides at a heterozygous locus
of interest on chromosome 21 will be present in a ratio of 0.20 (40
for fetal G allele/(60*2 for maternal A allele+40*2 for fetal A
allele). By contrast, the ratio for chromosome 1 will remain at
0.25, and thus the 5% difference in ratios will signify an
additional chromosome. One to tens to hundreds to thousands of loci
of interest can be analyzed.
[0368] In another embodiment, the loci of interest on the template
DNA from the sample from the pregnant female can be genotyped
without prior identification of the homozygous maternal loci of
interest. It is not necessary to genotype the maternal template DNA
prior to analysis of the template DNA containing both maternal and
fetal template. DNA.
[0369] The ratio of the alleles at the loci of interest can be used
to determine the presence or absence of a chromosomal abnormality.
The template DNA from the sample from the pregnant female contains
both maternal template DNA and fetal template DNA. There are 3
possibilities at each SNP for either the maternal template DNA or
the fetal template DNA: heterozygous, homozygous for allele 1, or
homozygous for allele 2. The possible nucleotide ratios for a SNP
that is either an adenine or a guanine are shown in Table II. The
ratios presented in Table II are calculated with the fetal DNA at
50% of the DNA in the sample from the pregnant female.
TABLE-US-00026 TABLE II Ratios for nucleotides for a heterozygous
SNP. Fetal SNP Maternal SNP A/A G/G A/G A/A 100% A N/A 75% A, 25% G
G/G N/A 100% G 25% A, 75% G A/G 75% A, 25% G 25% A, 75% G 50% A,
50% G
[0370] There are three nucleotide ratios: 100% of a single
nucleotide, 50:50, or 75:25. These ratios will vary depending on
the amount of fetal DNA present in sample from the pregnant female.
However, the percentage of fetal DNA should be constant regardless
of the chromosome analyzed. Therefore, if chromosomes are present
in two copies, the above calculated ratios will be seen.
[0371] On the other hand, these percentages will vary when an
additional chromosome is present. For example, assume that SNP X
can be adenine or guanine, and that the percentage of fetal DNA in
the sample from the pregnant female is 50%. Analysis of the loci of
interest on chromosome 1 will provide the ratios discussed above:
100:0, 50:50, and 75:25. The possible ratios for a SNP that is A/G
with an additional chromosome are provided in Table III.
TABLE-US-00027 TABLE III Nucleotides ratios at a SNP when an
additional copy of a chromosome is present Maternal Fetal SNP SNPX
A/A/A G/G/G A/G/G A/A/G A/A 100% A N/A 60% A, 40% G 80% A, 20% G
G/G N/A 100% G 20% A, 80% G 40% A, 60% G A/G 80% A, 20% G 20% A,
80% G 40% A, 60% G 60% A, 40% G
[0372] The possible ratios for the alleles at a heterozygous SNP
with an additional copy of a chromosome are: 0:100, 40:60, and
20:80. Two of these ratios, 40:60, and 20:80 differ from the ratios
of alleles at heterozygous SNPs obtained with two copies of a
chromosome. As discussed above, the ratios for the nucleotides at a
heterozygous SNP depend on the amount of fetal DNA present in the
sample. However, the ratios, whatever they are, will remain
constant across chromosomes unless there is a chromosomal
abnormality.
[0373] The ratio of alleles at heterozygous loci of interest on a
chromosome can be compared to the ratio for alleles at heterozygous
loci of interest on a different chromosome. For example, the ratio
for multiple loci of interest on chromosome 1 (the ratio at SNP 1,
SNP 2, SNP 3, SNP 4, etc.) can be compared to the ratio for
multiple loci of interest on chromosome 21 (the ratio at SNP A, SNP
B, SNP C, SNP D, etc.). Any chromosome can be compared to any other
chromosome. There is no limit to the number of chromosomes that can
be compared.
[0374] Referring back to the data in Tables II and III, the ratios
for nucleotides at a heterozygous SNP on chromosome 1, which was
present in two copies, were 25:75, and 50:50. On the other, the
ratio for nucleotides at a heterozygous SNP on chromosome 21, which
was present in three copies, were 40:60, and 20:80. The difference
between these two ratios indicates a chromosomal abnormality. The
ratios can be pre-calculated for the full range of varying degrees
of fetal DNA present in the maternal serum. Tables II and III
demonstrate that both maternal homozygous and heterozygous loci of
interest can be used to detect the presence of a fetal chromosomal
abnormality.
[0375] The above example illustrates how the ratios for nucleotides
at heterozygous SNPs can be used to detect the presence of an
additional chromosome. The same type of analysis can be used to
detect chromosomal rearrangements, translocations,
mini-chromosomes, duplications of regions of chromosomes,
monosomies, deletions of regions of chromosomes, and fragments of
chromosomes. The method does not require genotyping of the mother
or the father, however, it may be done to reduce the number of SNPs
that need to be analyzed with the plasma sample.
[0376] The present invention does not quantitate the amount of a
fetal gene product, nor is the utility of the present invention
limited to the analysis of genes found on the Y chromosome. The
present invention does not merely rely on the detection of a
paternally inherited nucleic acid, rather, the present invention
provides a method that allows the ratio of maternal to fetal
alleles at loci of interest, including SNPs, to be calculated.
[0377] In another embodiment, a single allele at a locus of
interest can be used to determine the presence or absence of a
chromosomal abnormality and detect a genetic disorder in the fetus.
In a preferred embodiment, the maternal allele at a locus of
interest is used to determine the presence or absence of a
chromosomal abnormality in the fetus. The biological mother can be
genotyped to identify a homozygous locus of interest. Likewise, the
biological father can be genotyped to identify a homozygous locus
of interest. The locus of interest wherein the maternal template
DNA is homozygous for one allele and the paternal template DNA is
homozygous for the other allele is analyzed using the template DNA
obtained from the plasma of the mother, which contains both
maternal and fetal template DNA. Any number of loci of interest can
be analyzed including but not limited to 1, 1-5, 5-10, 10-20,
20-30, 30-40, 40-50, 50-60, 60-70, 70-80, 80-90, 90-100, 100-150,
150-200, 200-250, 250-300, 300-500, 500-1000, 1000-2000, 2000-3000,
3000-4000, 4000-8000, 8000-16000, 16000-32000 or greater than 32000
loci of interest.
[0378] In a preferred embodiment, the signal from the maternal
genome and the fetal allele, which was inherited from the mother,
at the locus of interest is quantitated. For example, if the 5'
overhang, which is generated after digestion with the type IIS
enzyme, is filled in with a nucleotide that is fluorescently
labeled, the intensity of the incorporated dye can be
quantitated
Maternal Template DNA--Homozygous for Adenine
TABLE-US-00028 [0379] Allele 1 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4 Allele 2 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4
Paternal Template DNA--Homozygous for Cytosine
TABLE-US-00029 [0380] Allele 1 5' CCG C* 3' GGC G G T G Overhang
position 1 2 3 4 Allele 2 5' CCG C* 3' GGC T G T G Overhang
position 1 2 3 4
Template DNA in the Plasma--Both Maternal Template DNA and Fetal
Template DNA
[0381] Maternal Template DNA--Homozygous for Adenine
TABLE-US-00030 [0381] Allele 1 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4 Allele 2 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4
Fetal Template DNA--Heterozygous
TABLE-US-00031 [0382] Allele 1 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4 Allele 2 5' CCG ddC 3' GGC T G T G Overhang
position 1 2 3 4
[0383] The template DNA obtained from the plasma of the pregnant
female is filled in with labeled ddATP, and unlabeled ddCTP
(depicted as ddC above), ddGTP, and ddTTP. The plasma DNA contains
two maternal adenine alleles, and one fetal adenine allele. By
filling in with labeled ddATP and unlabeled ddCTP, only the
maternal allele and the fetal allele inherited from the mother are
detected. The paternal allele is not detected in this manner. The
fill-in reactions can be performed as described in the Examples
below.
[0384] A single locus of interest can be analyzed or multiple loci
of interest. The intensity of the maternal allele at multiple loci
of interest can be quantitated. An average can be calculated for a
chromosome and compared to the average obtained for a different
chromosome. For example, the average intensity of the maternal
allele and the fetal allele inherited from the mother at chromosome
1 can be compared to the average intensity of the maternal allele
and the fetal allele inherited from the mother at chromosomes 13,
18, or 21. In a preferred embodiment, chromosomes 13, 15, 18, 21,
22, X and Y, when applicable, are compared.
[0385] The signal from a locus of interest may be stronger than
another locus of interest. However, there is no reason why the
signal from the locus of interest on one chromosome would be
stronger than the signal from the locus of interest on another
chromosome. While the signal from various loci of interest may be
variable, the variation should be seen across the genome. The
average signal of the loci of interest should be the same when any
chromosomes are compared.
[0386] The conditions of the PCR reaction can be optimized so that
an equivalent amount of PCR product is produced. For example, the
concentration of the primers, the concentration of nucleotides, and
the number of cycles for each loci of interest can be optimized. In
addition, the fill-in reactions can be done under conditions such
that any increase in a specific allele can be detected. The fill-in
reaction conditions can be optimized to detect any increase in the
allele of interest including but not limited to the concentration
of reagents, the time of the fill-in reaction, and the temperature
of the reaction.
[0387] With a normal genetic karyotype, the signal at each locus of
interest comprise signal from the maternal genome, and signal from
the fetal allele, which was inherited from the mother. The percent
of fetal DNA in the sample remains constant, regardless of the
chromosome that is analyzed. For example, if at SNP X, the maternal
genome is A/A, and the paternal genome is G/G, then the fetal
genome will be A/G, and the fetal adenine allele will comprise a
specified percentage of the signal from the adenine allele. If the
percentage of fetal DNA is 20% in the maternal plasma, then the
fetal adenine allele will contribute 20% of the signal for the
adenine allele. The contribution of the fetal allele, which was
inherited from the mother, will be constant for any locus of
interest that is analyzed.
[0388] When there is a chromosomal abnormality, the signal from the
maternal genome and the fetal allele, which was inherited from the
mother, at the loci of interest will differ from the signal
observed for other chromosomes. For example, with a Trisomy, the
signal at the locus of interest will comprise the maternal genome
and two fetal alleles, which were inherited from the mother. The
signal from the loci of interest for the chromosome that is present
in three copies will have the contribution of an additional fetal
allele, which will alter the signal of the alleles at these loci of
interest.
[0389] In another embodiment, a ratio can be calculated using a
single allele and a standard DNA of known quantity. In a preferred
embodiment, a ratio is calculated using the alleles of the maternal
genome, and the fetal allele, which was inherited from the mother,
and a standard DNA. The biological mother can be genotyped to
identify a homozygous locus of interest. Likewise, the biological
father can be genotyped to identify a homozygous locus of interest.
The locus of interest wherein the maternal template DNA is
homozygous for one allele and the paternal template DNA is
homozygous for the other allele is analyzed using the template DNA
obtained from the plasma of the mother, which contains both
maternal and fetal template DNA.
[0390] In a preferred embodiment, the signal from the maternal
genome and the fetal allele, which was inherited from the mother,
at the locus of interest is quantitated. For example, if the 5'
overhang, which is generated after digestion with the type IIS
enzyme, is filled in with a nucleotide that is fluorescently
labeled, the intensity of the incorporated dye can be
quantitated.
Template DNA in the Plasma--Both Maternal Template DNA and Fetal
Template DNA
[0391] Maternal Template DNA--Homozygous for Adenine
TABLE-US-00032 [0391] Allele 1 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4 Allele 2 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4
Fetal Template DNA--Heterozygous
TABLE-US-00033 [0392] Allele 1 5' CCG A* 3' GGC T G T G Overhang
position 1 2 3 4 Allele 2 5' CCG ddC 3' GGC T G T G Overhang
position 1 2 3 4
[0393] The template DNA obtained from the plasma of the pregnant
female is filled in with labeled ddATP, and unlabeled ddCTP
(depicted as ddC above), ddGTP, and ddTTP. The plasma DNA contains
two maternal adenine alleles, and one fetal adenine allele. By
filling in with labeled ddATP and unlabeled ddCTP, only the
maternal allele and the fetal allele inherited from the mother are
detected.
[0394] A single locus of interest or multiple loci of interest can
be analyzed. For each locus of interest, a DNA molecule is designed
to migrate at about the same position as the locus of interest. In
a preferred embodiment, the DNA molecule is of known quantity. A
ratio is calculated using the alleles of the maternal genome and
the fetal allele, which was inherited from the mother, and the DNA
molecule designed to migrate at about the same position as the
locus of interest. For example, if the locus of interest is
designed to migrate at 30 base pairs, the DNA molecule can be
designed to migrate at about 30 base pairs including but not
limited to 20-25, 25-30, 30-35, 35-45, and greater than 45. The
alleles of the maternal genome and the fetal allele, which was
inherited from the mother, and the standard DNA molecule can be
analyzed in the same reaction or can be analyzed in a separate
reaction. The alleles of the maternal genome and the fetal allele,
which was inherited from the mother, and the standard DNA molecule
can be analyzed in the same lane of a gel or can be analyzed in
separate lanes of a gel. The use of standard DNA molecules of known
quantity, which are designed to migrate at the same position as the
loci of interest, will correct for various factors including but
not limited to the intensity of the bands relative to the location
on the gel.
[0395] The ratio of multiple loci of interest on a chromosome can
be quantitated, and an average calculated. The average can be
compared to the average obtained for another chromosome. The ratio
is used to indicate the presence or absence of a chromosomal
abnormality. Analysis of the alleles of the maternal genome and the
fetal allele also allows detection of single gene or multi-gene
genetic disorders.
[0396] Any chromosome of any organism can be analyzed using the
methods of the invention. For example, in humans, chromosome 1, 2,
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
21, 22, X or Y can be analyzed using the methods of the invention.
The ratio for the alleles at a heterozygous locus of interest on
any chromosome can be compared to the ratio for the alleles at a
heterozygous locus of interest on any other chromosome.
[0397] Thus, the present invention provides a non-invasive
technique, which is independent of fetal cell isolation, for rapid,
accurate and definitive detection of chromosome abnormalities in a
fetus. The present invention also provides a non-invasive method
for determining the sequence of DNA from a fetus. The present
invention can be used to detect any alternation in gene sequence as
compared to the wild type sequence including but not limited to
point mutation, reading frame shift, transition, transversion,
addition, insertion, deletion, addition-deletion, frame-shift,
missense, reverse mutation, and microsatellite alteration.
Detection of Fetal Chromosomal Abnormalities Using Short Tandem
Repeats
[0398] Short tandem repeats (STRs) are short sequences of DNA,
normally of 2-5 base pairs in length, which are repeated numerous
times in a head-tail manner. Tandemly repeated DNA sequences are
widespread throughout the human genome, and show sufficient
variability among the individuals in a population. Minisatellites
have core repeats with 9-80 base pairs.
[0399] In another embodiment, short tandem repeats can be used to
detect fetal chromosomal abnormalities. Template DNA can be
obtained from a nucleic acid containing sample including but not
limited to cell, tissue, blood, serum, plasma, saliva, urine,
tears, vaginal secretion, lymph fluid, cerebrospinal fluid, mucosa
secretion, peritoneal fluid, ascitic fluid, fecal matter, or body
exudates. In another embodiment, a cell lysis inhibitor is added to
the nucleic acid containing sample. In a preferred embodiment, the
template DNA is obtained from the blood of a pregnant female. In
another embodiment, the template DNA is obtained from the plasma or
serum from the blood of a pregnant female.
[0400] The template DNA obtained from the blood of the pregnant
female will contain both fetal DNA and maternal DNA. The fetal DNA
comprises STRs from the mother and the father. The variation in the
STRs between the mother and father can be used to detect
chromosomal abnormalities.
[0401] Primers can be designed to amplify short tandem repeats. Any
method of amplification can be used including but not limited to
polymerase chain reaction, self-sustained sequence reaction, ligase
chain reaction, rapid amplification of cDNA ends, polymerase chain
reaction and ligase chain reaction, Q-beta phage amplification,
strand displacement amplification, and splice overlap extension
polymerase chain reaction. In a preferred embodiment, PCR is
used.
[0402] Any number of short tandem repeats can be analyzed including
but not limited to 1-5, 5-10, 10-50, 50-100, 100-200, 200-300,
300-400, 400-500, 500-1000, and greater than 1000. The short tandem
repeats can be analyzed in a single PCR reaction or in multiple PCR
reactions. In a preferred embodiment, STRs from multiple
chromosomes are analyzed.
[0403] After amplification, the PCR products can be analyzed by any
number of methods including but not restricted to get
electrophoresis, and mass spectrometry. The template DNA from the
pregnant female comprises STRs of maternal and paternal origin. The
STRs of paternal origin represent the fetal DNA. The paternal and
maternal STRs may be identical in length or the maternal and the
paternal STRs may differ.
[0404] Heterozygous STRs are those of which the maternal and
paternal differ in length. The amount of each PCR product can be
quantitated for each heterozygous STR. With a normal number of
chromosomes, the amount of each PCR product should be approximately
equal. However, with an extra chromosome, one of the STR PCR
products will be present at a greater amount.
[0405] For example, multiple STRs on chromosome 1 can be analyzed
on the template DNA obtained from the blood of the pregnant female.
Each STR, whether of maternal or paternal origin, should be present
at approximately the same amount. Likewise, with two chromosome
21s, each STR should be present at approximately the same amount.
However, with a trisomy 21, one of the STR PCR products, when the
maternal and paternal differ in length (a heterozygous STR) should
be present at a higher amount. The ratio for each heterozygous STR
on one chromosome can be compared to the ratio for each
heterozygous STR on a different chromosome, wherein a difference
indicates the presence or absence of a chromosomal abnormality.
Kits
[0406] The methods of the invention are most conveniently practiced
by providing the reagents used in the methods in the form of kits.
A kit preferably contains one or more of the following components:
written instructions for the use of the kit, appropriate buffers,
salts, DNA extraction detergents, primers, nucleotides, labeled
nucleotides, 5' end modification materials, and if desired, water
of the appropriate purity, confined in separate containers or
packages, such components allowing the user of the kit to extract
the appropriate nucleic acid sample, and analyze the same according
to the methods of the invention. The primers that are provided with
the kit will vary, depending upon the purpose of the kit and the
DNA that is desired to be tested using the kit.
[0407] A kit can also be designed to detect a desired or variety of
single nucleotide polymorphisms, especially those associated with
an undesired condition or disease. For example, one kit can
comprise, among other components, a set or sets of primers to
amplify one or more loci of interest associated with Huntington's
disease. Another kit can comprise, among other components, a set or
sets of primers for genes associated with a predisposition to
develop type I or type II diabetes. Still, another kit can
comprise, among other components, a set or sets of primers for
genes associated with a predisposition to develop heart disease.
Details of utilities for such kits are provided in the "Utilities"
section below.
Utilities
[0408] The methods of the invention can be used whenever it is
desired to know the genotype of an individual. The method of the
invention is especially useful for the detection of genetic
disorders. The method of the invention is especially useful as a
non-invasive technique for the detection of genetic disorders in a
fetus. In a preferred embodiment, the method of the invention
provides a method for identification of single nucleotide
polymorphisms.
[0409] In a preferred embodiment, the method is useful for
detecting chromosomal abnormalities including but not limited to
trisomies, monosomies, duplications, deletions, additions,
chromosomal rearrangements, translocations, and other aneuploidies.
The method is especially useful for the detection of chromosomal
abnormalities in a fetus.
[0410] In a preferred embodiment, the method of the invention
provides a method for identification of the presence of a disease
in a fetus, especially a genetic disease that arises as a result of
the presence of a genomic sequence, or other biological condition
that it is desired to identify in an individual for which it is
desired to know the same. The identification of such sequence in
the fetus based on the presence of such genomic sequence can be
used, for example, to determine if the fetus is a carrier or to
assess if the fetus is predisposed to developing a certain genetic
trait, condition or disease. The method of the invention is
especially useful in prenatal genetic testing of parents and
child.
[0411] Examples of diseases that can be diagnosed by this invention
are listed in Table IV.
TABLE-US-00034 TABLE IV Achondroplasia Adrenoleukodystrophy,
X-Linked Agammaglobulinemia, X-Linked Alagille Syndrome
Alpha-Thalassemia X-Linked Mental Retardation Syndrome Alzheimer
Disease Alzheimer Disease, Early-Onset Familial Amyotrophic Lateral
Sclerosis Overview Androgen Insensitivity Syndrome Angelman
Syndrome Ataxia Overview, Hereditary Ataxia-Telangiectasia Becker
Muscular Dystrophy also The Dystrophinopathies) Beckwith-Wiedemann
Syndrome Beta-Thalassemia Biotinidase Deficiency Branchiootorenal
Syndrome BRCA1 and BRCA2 Hereditary Breast/Ovarian Cancer Breast
Cancer CADASIL Canavan Disease Cancer Charcot-Marie-Tooth
Hereditary Neuropathy Charcot-Marie-Tooth Neuropathy Type 1
Charcot-Marie-Tooth Neuropathy Type 2 Charcot-Marie-Tooth
Neuropathy Type 4 Charcot-Marie-Tooth Neuropathy Type X Cockayne
Syndrome Colon Cancer Contractural Arachnodactyly, Congenital
Craniosynostosis Syndromes (FGFR-Related) Cystic Fibrosis
Cystinosis Deafness and Hereditary Hearing Loss DRPLA
(Dentatorubral-Pallidoluysian Atrophy) DiGeorge Syndrome (also
22q11 Deletion Syndrome) Dilated Cardiomyopathy, X-Linked Down
Syndrome (Trisomy 21) Duchenne Muscular Dystrophy (also The
Dystrophinopathies) Dystonia, Early-Onset Primary (DYT1)
Dystrophinopathies, The Ehlers-Danlos Syndrome, Kyphoscoliotic Form
Ehlers-Danlos Syndrome, Vascular Type Epidermolysis Bullosa Simplex
Exostoses, Hereditary Multiple Facioscapulohumeral Muscular
Dystrophy Factor V Leiden Thrombophilia Familial Adenomatous
Polyposis (FAP) Familial Mediterranean Fever Fragile X Syndrome
Friedreich Ataxia Frontotemporal Dementia with Parkinsonism-17
Galactosemia Gaucher Disease Hemochromatosis, Hereditary Hemophilia
A Hemophilia B Hemorrhagic Telangiectasia, Hereditary Hearing Loss
and Deafness, Nonsyndromic, DFNA (Connexin 26) Hearing Loss and
Deafness, Nonsyndromic, DFNB 1 (Connexin 26) Hereditary Spastic
Paraplegia Hermansky-Pudlak Syndrome Hexasaminidase A Deficiency
(also Tay-Sachs) Huntington Disease Hypochondroplasia Ichthyosis,
Congenital, Autosomal Recessive Incontinentia Pigmenti Kennedy
Disease (also Spinal and Bulbar Muscular Atrophy) Krabbe Disease
Leber Hereditary Optic Neuropathy Lesch-Nyhan Syndrome Leukemias
Li-Fraumeni Syndrome Limb-Girdle Muscular Dystrophy Lipoprotein
Lipase Deficiency, Familial Lissencephaly Marfan Syndrome MELAS
(Mitochondrial Encephalomyopathy, Lactic Acidosis, and Stroke-Like
Episodes) Monosomies Multiple Endocrine Neoplasia Type 2 Multiple
Exostoses, Hereditary Muscular Dystrophy, Congenital Myotonic
Dystrophy Nephrogenic Diabetes Insipidus Neurofibromatosis 1
Neurofibromatosis 2 Neuropathy with Liability to Pressure Palsies,
Hereditary Niemann-Pick Disease Type C Nijmegen Breakage Syndrome
Norrie Disease Oculocutaneous Albinism Type 1 Oculopharyngeal
Muscular Dystrophy Ovarian Cancer Pallister-Hall Syndrome Parkin
Type of Juvenile Parkinson Disease Pelizaeus-Merzbacher Disease
Pendred Syndrome Peutz-Jeghers Syndrome Phenylalanine Hydroxylase
Deficiency Prader-Willi Syndrome PROP 1-Related Combined Pituitary
Hormone Deficiency (CPHD) Prostate Cancer Retinitis Pigmentosa
Retinoblastoma Rothmund-Thomson Syndrome Smith-Lemli-Opitz Syndrome
Spastic Paraplegia, Hereditary Spinal and Bulbar Muscular Atrophy
(also Kennedy Disease) Spinal Muscular Atrophy Spinocerebellar
Ataxia Type 1 Spinocerebellar Ataxia Type 2 Spinocerebellar Ataxia
Type 3 Spinocerebellar Ataxia Type 6 Spinocerebellar Ataxia Type 7
Stickler Syndrome (Hereditary Arthroophthalmopathy) Tay-Sachs (also
GM2 Gangliosidoses) Trisomies Tuberous Sclerosis Complex Usher
Syndrome Type I Usher Syndrome Type II Velocardiofacial Syndrome
(also 22q11 Deletion Syndrome) Von Hippel-Lindau Syndrome Williams
Syndrome Wilson Disease X-Linked Adrenoleukodystrophy X-Linked
Agammaglobulinemia X-Linked Dilated Cardiomyopathy (also The
Dystrophinopathies) X-Linked Hypotonic Facies Mental Retardation
Syndrome
[0412] The method of the invention is useful for screening an
individual at multiple loci of interest, such as tens, hundreds, or
even thousands of loci of interest associated with a genetic trait
or genetic disease by sequencing the loci of interest that are
associated with the trait or disease state, especially those most
frequently associated with such trait or condition. The invention
is useful for analyzing a particular set of diseases including but
not limited to heart disease, cancer, endocrine disorders, immune
disorders, neurological disorders, musculoskeletal disorders,
ophthalmologic disorders, genetic abnormalities, trisomies,
monosomies, transversions, translocations, skin disorders, and
familial diseases.
[0413] The method of the invention can also be used to confirm or
identify the relationship of a DNA of unknown sequence to a DNA of
known origin or sequence, for example, for use in, maternity or
paternity testing, and the like.
[0414] Having now generally described the invention, the same will
become better understood by reference to certain specific examples
which are included herein for purposes of illustration only and are
not intended to be limiting unless otherwise specified.
EXAMPLES
[0415] The following examples are illustrative only and are not
intended to limit the scope of the invention as defined by the
claims.
Example 1
[0416] DNA sequences were amplified by PCR, wherein the annealing
step in cycle 1 was performed at a specified temperature, and then
increased in cycle 2, and further increased in cycle 3 for the
purpose of reducing non-specific amplification. The TM1 of cycle 1
of PCR was determined by calculating the melting temperature of the
3' region, which anneals to the template DNA, of the second primer.
For example, in FIG. 1B, the TM1 can be about the melting
temperature of region "c." The annealing temperature was raised in
cycle 2, to TM2, which was about the melting temperature of the 3'
region, which anneals to the template DNA, of the first primer. For
example, in FIG. 1C, the annealing temperature (TM2) corresponds to
the melting temperature of region "b." In cycle 3, the annealing
temperature was raised to TM3, which was about the melting
temperature of the entire sequence of the second primer. For
example, in FIG. 1D, the annealing temperature (TM3) corresponds to
the melting temperature of region "c"+region "d". The remaining
cycles of amplification were performed at TM3.
Preparation of Template DNA
[0417] The template DNA was prepared from a 5 ml sample of blood
obtained by venipuncture from a human volunteer with informed
consent. Blood was collected from 36 volunteers. Template DNA was
isolated from each blood sample using QIAmp DNA Blood Midi Kit
supplied by QIAGEN (Catalog number 51183). Following isolation, the
template DNA from each of the 36 volunteers was pooled for further
analysis.
Primer Design
[0418] The following four single nucleotide polymorphisms were
analyzed: SNP HC21S00340, identification number as assigned by
Human Chromosome 21 cSNP Database, (FIG. 3, lane 1) located on
chromosome 21; SNP TSC 0095512 (FIG. 3, lane 2) located on
chromosome 1, SNP TSC 0214366 (FIG. 3, lane 3) located on
chromosome 1; and SNP TSC 0087315 (FIG. 3, lane 4) located on
chromosome 1. The SNP Consortium Ltd database can be accessed at
http://snp.cshl.org/, website address effective as of Feb. 14,
2002.
[0419] SNP HC21S00340 was amplified using the following
primers:
TABLE-US-00035 First primer: (SEQ ID NO: 9)
5'TAGAATAGCACTGAATTCAGGAATACAATCATTGTCAC 3' Second primer: (SEQ ID
NO: 10) 5'ATCACGATAAACGGCCAAACTCAGGTTA3'
[0420] SNP TSC0095512 was amplified using the following
primers:
TABLE-US-00036 First primer: (SEQ ID NO: 11)
5'AAGTTTAGATCAGAATTCGTGAAAGCAGAAGTTGTCTG 3' Second primer: (SEQ ID
NO: 12) 5'TCTCCAACTAACGGCTCATCGAGTAAAG 3'
[0421] SNP TSC0214366 was amplified using the following
primers:
TABLE-US-00037 First primer: (SEQ ID NO: 13)
5'ATGACTAGCTATGAATTCGTTCAAGGTAGAAAATGGAA 3' Second primer: (SEQ ID
NO: 14) 5'GAGAATTAGAACGGCCCAAATCCCACTC3'
[0422] SNP TSC 0087315 was amplified using the following
primers:
TABLE-US-00038 First primer: (SEQ ID NO: 15)
5'TTACAATGCATGAATTCATCTTGGTCTCTCAAAGTGC 3' Second primer: (SEQ ID
NO: 16) 5'TGGACCATAAACGGCCAAAAACTGTAAG 3'.
[0423] All primers were designed such that the 3' region was
complementary to either the upstream or downstream sequence
flanking each locus of interest and the 5' region contained a
restriction enzyme recognition site. The first primer contained a
biotin tag at the 5' end and a recognition site for the restriction
enzyme EcoRI. The second primer contained the recognition site for
the restriction enzyme BceA I.
PCR Reaction
[0424] All four loci of interest were amplified from the template
genomic DNA using PCR (U.S. Pat. Nos. 4,683,195 and 4,683,202). The
components of the PCR reaction were as follows: 40 ng of template
DNA, 5 .mu.M first primer, 5 .mu.M second primer, 1.times.
HotStarTaq Master Mix as obtained from Qiagen (Catalog No. 203443).
The HotStarTaq Master Mix contained DNA polymerase, PCR buffer, 200
.mu.M of each dNTP, and 1.5 mM MgCl.sub.2.
[0425] Amplification of each template DNA that contained the SNP of
interest was performed using three different series of annealing
temperatures, herein referred to as low stringency annealing
temperature, medium stringency annealing temperature, and high
stringency annealing temperature. Regardless of the annealing
temperature protocol, each PCR reaction consisted of 40 cycles of
amplification. PCR reactions were performed using the HotStarTaq
Master Mix Kit supplied by QIAGEN. As instructed by the
manufacturer, the reactions were incubated at 95.degree. C. for 15
min. prior to the first cycle of PCR. The denaturation step after
each extension step was performed at 95.degree. C. for 30 sec. The
annealing reaction was performed at a temperature that permitted
efficient extension without any increase in temperature.
[0426] The low stringency annealing reaction comprised three
different annealing temperatures in each of the first three cycles.
The annealing temperature for the first cycle was 37.degree. C. for
30 sec.; the annealing temperature for the second cycle was
57.degree. C. for 30 sec.; the annealing temperature for the third
cycle was 64.degree. C. for 30 sec. Annealing was performed at
64.degree. C. for subsequent cycles until completion.
[0427] As shown in the photograph of the gel (FIG. 3A), multiple
bands were observed after amplification of SNP TSC 0087315 (lane
4). Amplification of SNP HC21S00340 (lane 1), SNP TSC0095512 (lane
2), and SNP TSC0214366 (lane 3) generated a single band of high
intensity and one band of faint intensity, which was of higher
molecular weight. When the low annealing temperature conditions
were used, the correct size product was generated and this was the
predominant product in each reaction.
[0428] The medium stringency annealing reaction comprised three
different annealing temperatures in each of the first three cycles.
The annealing temperature for the first cycle was 40.degree. C. for
30 seconds; the annealing temperature for the second cycle was
60.degree. C. for 30 seconds; and the annealing temperature for the
third cycle was 67.degree. C. for 30 seconds. Annealing was
performed at 67.degree. C. for subsequent cycles until completion.
Similar to what was observed under low stringency annealing
conditions, amplification of SNP TSC0087315 (FIG. 3B, lane 4)
generated multiple bands under conditions of medium stringency.
Amplification of the other three SNPs (lanes 1-3) produced a single
band. These results demonstrate that variable annealing
temperatures can be used to cleanly amplify loci of interest from
genomic DNA with a primer that has an annealing length of 13
bases.
[0429] The high stringency annealing reaction was comprised of
three different annealing temperatures in each of the first three
cycles. The annealing temperature of the first cycle was 46.degree.
C. for 30 seconds; the annealing temperature of the second cycle
was 65.degree. C. for 30 seconds; and the annealing temperature for
the third cycle was 72.degree. C. for 30 seconds. Annealing was
performed at 72.degree. C. for subsequent cycles until completion.
As shown in the photograph of the gel (FIG. 3C), amplification of
SNP TSC0087315 (lane 4) using the high stringency annealing
temperatures generated a single band of the correct molecular
weight. By raising the annealing temperatures for each of the first
three cycles, non-specific amplification was eliminated.
Amplification of SNP TSC0095512 (lane 2) generated a single band.
SNPs HC21S00340 (lane 1), and TSC0214366 (lane 3) failed to amplify
at the high stringency annealing temperatures, however, at the
medium stringency annealing temperatures, these SNPs amplified as a
single band. These results demonstrate that variable annealing
temperatures can be used to reduce non-specific PCR products, as
demonstrated for SNP TSC0087315 (FIG. 3, lane 4).
Example 2
[0430] SNPs on chromosomes 1 (TSC0095512), 13 (TSC0264580), and 21
(HC21S00027) were analyzed. SNP TSC0095512 was analyzed using two
different sets of primers, and SNP HC21S00027 was analyzed using
two types of reactions for the incorporation of nucleotides.
Preparation of Template DNA
[0431] The template DNA was prepared from a 5 ml sample of blood
obtained by venipuncture from a human volunteer with informed
consent. Template DNA was isolated using the QIAmp DNA Blood Midi
Kit supplied by QIAGEN (Catalog number 51183). The template DNA was
isolated as per instructions included in the kit. Following
isolation, template DNA from thirty-six human volunteers were
pooled together and cut with the restriction enzyme EcoRI. The
restriction enzyme digestion was performed as per manufacturer's
instructions.
Primer Design
[0432] SNP HC21S00027 was amplified by PCR using the following
primer set:
TABLE-US-00039 First primer: (SEQ ID NO: 17) 5'
ATAACCGTATGCGAATTCTATAATTTTCCTGATAAAGG 3' Second primer: (SEQ ID
NO: 18) 5' CTTAAATCAGGGGACTAGGTAAACTTCA 3'.
[0433] The first primer contained a biotin tag at the extreme 5'
end, and the nucleotide sequence for the restriction enzyme EcoRI.
The second primer contained the nucleotide sequence for the
restriction enzyme BsmF I (FIG. 4A).
[0434] Also, SNP HC21S00027 was amplified by PCR using the same
first primer but a different second primer with the following
sequence:
TABLE-US-00040 Second primer: 5' CTTAAATCAGACGGCTAGGTAAACTTCA 3'
(SEQ ID NO: 19)
[0435] This second primer contained the recognition site for the
restriction enzyme BceA I (FIG. 4B).
[0436] SNP TSC0095512 was amplified by PCR using the following
primers:
TABLE-US-00041 First primer: (SEQ ID NO: 11) 5'
AAGTTTAGATCAGAATTCGTGAAAGCAGAAGTTGTCTG 3' Second primer: (SEQ ID
NO: 20) 5' TCTCCAACTAGGGACTCATCGAGTAAAG 3'.
[0437] The first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI. The
second primer contained a restriction enzyme recognition site for
BsmF I (FIG. 4C).
[0438] Also, SNP TSC0095512 was amplified using the same first
primer and a different second primer with the following
sequence:
TABLE-US-00042 Second primer: 5'TCTCCAACTAACGGCTCATCGAGTAAAG 3'
(SEQ ID NO: 12)
[0439] This second primer contained the recognition site for the
restriction enzyme BceA I (FIG. 4D).
[0440] SNP TSC0264580, which is located on chromosome 13, was
amplified with the following primers:
TABLE-US-00043 First primer: (SEQ ID NO: 21) 5'
AACGCCGGGCGAGAATTCAGTTTTTCAACTTGCAAGG 3' Second primer: (SEQ ID NO:
22) 5' CTACACATATCTGGGACGTTGGCCATCC 3'.
[0441] The first primer contained a biotin tag at the extreme 5'
end and had a restriction enzyme recognition site for EcoRI. The
second primer contained a restriction enzyme recognition site for
BsmF I.
PCR Reaction
[0442] All loci of interest were amplified from the template
genomic DNA using the polymerase chain reaction (PCR, U.S. Pat.
Nos. 4,683,195 and 4,683,202, incorporated herein by reference). In
this example, the loci of interest were amplified in separate
reaction tubes but they could also be amplified together in a
single PCR reaction. For increased specificity, a "hot-start" PCR
was used. PCR reactions were performed using the HotStarTaq Master
Mix Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction can be optimized for each
locus of interest but in this example, 40 ng of template human
genomic DNA and 5 .mu.M of each primer were used. Forty cycles of
PCR were performed. The following PCR conditions were used:
[0443] (1) 95.degree. C. for 15 minutes and 15 seconds;
[0444] (2) 37.degree. C. for 30 seconds;
[0445] (3) 95.degree. C. for 30 seconds;
[0446] (4) 57.degree. C. for 30 seconds;
[0447] (5) 95.degree. C. for 30 seconds;
[0448] (6) 64.degree. C. for 30 seconds;
[0449] (7) 95.degree. C. for 30 seconds;
[0450] (8) Repeat steps 6 and 7 thirty nine (39) times;
[0451] (9) 72.degree. C. for 5 minutes.
[0452] In the first cycle of PCR, the annealing temperature was
about the inciting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[0453] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results. The PCR products for
SNP HC21S00027 and SNP TSC095512 are shown in FIGS. 5A-5D.
Purification of Fragment of Interest
[0454] The PCR products were separated from the genomic template
DNA. Each PCR product was divided into four separate reaction wells
of a Streptawell, transparent, High-Bind plate from Roche
Diagnostics GmbH (catalog number 1 645 692, as listed in Roche
Molecular Biochemicals, 2001 Biochemicals Catalog). The first
primers contained a 5' biotin tag so the PCR products bound to the
Streptavidin coated wells while the genomic template DNA did not.
The streptavidin binding reaction was performed using a Thermomixer
(Eppendorf) at 1000 rpm for 20 min. at 37.degree. C. Each well was
aspirated to remove unbound material, and washed three times with
1.times.PBS, with gentle mixing (Kandpal et al., Nucl. Acids Res.
18:1789-1795 (1990); Kaneoka et al., Biotechniques 10:30-34 (1991);
Green et al., Nucl. Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[0455] The purified PCR products were digested with the restriction
enzyme that bound the recognition site incorporated into the PCR
products from the second primer. SNP HC21 S00027 (FIGS. 6A and 68)
and SNP TSC0095512 (FIGS. 6C and 6D) were amplified in separate
reactions using two different second primers. FIG. 6A (SNP
HC21S00027) and FIG. 6C (SNP TSC0095512) depict the PCR products
after digestion with the restriction enzyme BsmF I (New England
Biolabs catalog number R0572S). FIG. 6B (SNP HC21S00027) and FIG.
6D (SNP TSC0095512) depict the PCR products after digestion with
the restriction enzyme BceA I (New England Biolabs, catalog number
R0623 S). The digests were performed in the Streptawells following
the instructions supplied with the restriction enzyme. SNP
TSC0264580 was digested with BsmF I. After digestion with the
appropriate restriction enzyme, the wells were washed three times
with PBS to remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[0456] The restriction enzyme digest described above yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[0457] For each SNP, four separate fill in reactions were
performed; each of the four reactions contained a different
fluorescently labeled dideoxynucleotide (ddATP, ddCTP, ddGTP, or
ddTTP). The following components were added to each fill in
reaction: 1 .mu.l of a fluorescently labeled dideoxynucleotide, 0.5
.mu.l of unlabeled ddNTPs (40 .mu.M), which contained all
nucleotides except the nucleotide that was fluorescently labeled, 2
.mu.l of 10.times. sequenase buffer, 0.25 .mu.l of Sequenase, and
water as needed for a 20 .mu.l reaction. All of the fill in
reactions were performed at 40.degree. C. for 10 min.
Non-fluorescently labeled nucleotides was purchased from Fermentas
Inc. (Hanover, Md.). All other labeling reagents were obtained from
Amersham (Thermo Sequenase Dye Terminator Cycle Sequencing Core
Kit, US 79565). In the presence of fluorescently labeled ddNTPs,
the 3' recessed end was extended by one base, which corresponds to
the SNP or locus of interest (FIG. 7A-7D).
[0458] A mixture of labeled ddNTPs and unlabeled dNTPs also was
used for the "fill in" reaction for SNP HC21S00027. The "fill in"
conditions were as described above except that a mixture containing
40 .mu.M unlabeled dNTPs, 1 .mu.l fluorescently labeled ddATP, I
fluorescently labeled ddCTP, 1 .mu.l fluorescently labeled ddGTP,
and 1 .mu.l ddTTP was used. The fluorescent ddNTPs were obtained
from Amersham (Thermo Sequenase Dye Terminator Cycle Sequencing
Core Kit, US 79565; Amersham did not publish the concentrations of
the fluorescent nucleotides). SNP HC21S00027 was digested with the
restriction enzyme BsmF I, which generated a 5' overhang of four
bases. As shown in FIG. 7E, if the first nucleotide incorporated is
a labeled dideoxynucleotide, the 3' recessed end is filled in by
one base, allowing detection of the SNP or locus of interest.
However, if the first nucleotide incorporated is a dNTP, the
polymerase continues to incorporate nucleotides until a ddNTP is
filled in. For example, the first two nucleotides can be filled in
with dNTPs, and the third nucleotide with a ddNTP, allowing
detection of the third nucleotide in the overhang. Thus, the
sequence of the entire 5' overhang can be determined, which
increases the information obtained from each SNP or locus of
interest.
[0459] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme (FIGS. 8A-8D). Digestion was
performed for 1 hour at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[0460] After release from the streptavidin matrix, 2-3 .mu.l of the
10 .mu.l sample was loaded in a 48 well membrane tray (The Gel
Company, catalog number TAM48-01). The sample in the tray was
absorbed with a 48 Flow Membrane Comb (The Gel Company, catalog
number AM48), and inserted into a 36 cm 5% acrylamide (urea) gel
(BioWhittaker Molecular Applications, Long Ranger Run Gel Packs,
catalog number 50691).
[0461] The sample was electrophoresed into the gel at 3000 volts
for 3 min. The membrane comb was removed, and the gel was run for 3
hours on an ABI 377 Automated Sequencing Machine. The incorporated
labeled nucleotide was detected by fluorescence.
[0462] As shown in FIG. 9A, from a sample of thirty six (36)
individuals, one of two nucleotides, either adenosine or guanine,
was detected at SNP HC21S00027. These are the two nucleotides
reported to exist at SNP HC21S00027
(http://snp.cshl.org/snpsearch.shtml).
[0463] One of two nucleotides, either guanine or cytosine, was
detected at SNP TS00095512 (FIG. 9B). The same results were
obtained whether the locus of interest was amplified with a second
primer that contained a recognition site for BceA I or the second
primer contained a recognition site for BsmF I.
[0464] As shown in FIG. 9C, one of two nucleotides was detected at
SNP TSC0264580, which was either adenosine or cytosine. These are
the two nucleotides reported for this SNP site
(http://snp.cshl.org/snpsearch.shtml). In addition, a thymidine was
detected one base from the locus of interest. In a sequence
dependent manner, BsmF I cuts some DNA molecules at the 10/14
position and other DNA molecules, which have the same sequence, at
the 11/15 position. When the restriction enzyme BsmF I cuts 11
nucleotides away on the sense strand and 15 nucleotides away on the
antisense strand, the 3' recessed end is one base from the SNP
site. The sequence of SNP TSC0264580 indicated that the base
immediately preceding the SNP site was a thymidine. The
incorporation of a labeled ddNTP into this position generated a
fragment one base smaller than the fragment that was cut at the
10/14 position. Thus, the DNA molecules cut at the 11/15 position
provided sequence information about the base immediately preceding
the SNP site, and the DNA molecules cut at the 10/14 position
provided sequence information about the SNP site.
[0465] SNP HC21S00027 was amplified using a second primer that
contained the recognition site for BsmF I. A mixture of labeled
ddNTPs and unlabeled dNTPs was used to fill in the 5' overhang
generated by digestion with BsmF I. If a dNTP was incorporated, the
polymerase continued to incorporate nucleotides until a ddNTP was
incorporated. A population of DNA fragments, each differing by one
base, was generated, which allowed the full sequence of the
overhang to be determined.
[0466] As seen in FIG. 9D, an adenosine was detected, which was
complementary to the nucleotide (a thymidine) immediately preceding
the SNP or locus of interest. This nucleotide was detected because
of the 11/15 cutting property of BsmF I, which is described in
detail above. A guanine and an adenosine were detected at the SNP
site, which are the two nucleotides reported for this SNP site
(FIG. 9A). The two nucleotides were detected at the SNP site
because the molecular weights of the dyes differ, which allowed
separation of the two nucleotides. The next nucleotide detected was
a thymidine, which is complementary to the nucleotide immediately
downstream of the SNP site. The next nucleotide detected was a
guanine, which was complementary to the nucleotide two bases
downstream of the SNP site. Finally, an adenosine was detected,
which was complementary to the third nucleotide downstream of the
SNP site. Sequence information was obtained not only for the SNP
site but for the nucleotide immediately preceding the SNP site and
the next three nucleotides.
[0467] None of the loci of interest contained a mutation. However,
if one of the loci of interest harbored a mutation including but
not limited to a point mutation, insertion, deletion, translocation
or any combination of said mutations, it could be identified by
comparison to the consensus or published sequence. Comparison of
the sequences attributed to each of the loci of interest to the
native, non-disease related sequence of the gene at each locus of
interest determines the presence or absence of a mutation in that
sequence. The finding of a mutation in the sequence is then
interpreted as the presence of the indicated disease, or a
predisposition to develop the same, as appropriate, in that
individual. The relative amounts of the mutated vs. normal or
non-mutated sequence can be assessed to determine if the subject
has one or two alleles of the mutated sequence, and thus whether
the subject is a carrier, or whether the indicated mutation results
in a dominant or recessive condition.
Example 3
[0468] Four loci of interest from chromosome 1 and two loci of
interest from chromosome 21 were amplified in separate PCR
reactions, pooled together, and analyzed. The primers were designed
so that each amplified locus of interest was a different size,
which allowed detection of the loci of interest.
Preparation of Template DNA
[0469] The template DNA was prepared from a 5 ml sample of blood
obtained by venipuncture from a human volunteer with informed
consent. Template DNA was isolated using the QIAmp DNA Blood Midi
Kit supplied by QIAGEN (Catalog number 51183). The template DNA was
isolated as per instructions included in the kit. Template DNA was
isolated from thirty-six human volunteers, and then pooled into a
single sample for further analysis.
Primer Design
[0470] SNP TSC 0087315 was amplified using the following
primers:
TABLE-US-00044 First primer: (SEQ ID NO: 15)
5'TTACAATGCATGAATTCATCTTGGTCTCTCAAAGTGC 3' Second primer: (SEQ ID
NO: 16) 5'TGGACCATAAACGGCCAAAAACTGTAAG3'.
[0471] SNP TSC0214366 was amplified using the following
primers:
TABLE-US-00045 First primer: (SEQ ID NO: 13)
5'ATGACTAGCTATGAATTCGTTCAAGGTAGAAAATGGAA 3' Second primer: (SEQ ID
NO: 14) 5'GAGAATTAGAACGGCCCAAATCCCACTC 3'
[0472] SNP TSC 0413944 was amplified with the following
primers:
TABLE-US-00046 First primer: (SEQ ID NO: 23) 5'
TACCTTTTGATCGAATTCAAGGCCAAAAATATTAAGTT 3' Second primer: (SEQ ID
NO: 24) 5' TCGAACTTTAACGGCCTTAGAGTAGAGA 3'
[0473] SNP TSC0095512 was amplified using the following
primers;
TABLE-US-00047 First primer: (SEQ ID NO: 11)
5'AAGTTTAGATCAGAATTCGTGAAAGCAGAAGTTGTCTG 3' Second primer: (SEQ ID
NO: 12) 5'TCTCCAACTAACGGCTCATCGAGTAAAG 3'
[0474] SNP HC21S00131 was amplified with the following primers:
TABLE-US-00048 First primer: (SEQ ID NO: 25) 5'
CGATTTCGATAAGAATTCAAAAGCAGTTCTTAGTTCAG 3' Second primer: (SEQ ID
NO: 26) 5'TGCGAATCTTACGGCTGCATCACATTCA 3'
[0475] SNP HC21S00027 was amplified with the following primers:
TABLE-US-00049 First primer: (SEQ ID NO: 17) 5'
ATAACCGTATGCGAATTCTATAATTTTCCTGATAAAGG 3' Second primer: (SEQ ID
NO: 19) 5' CTTAAATCAGACGGCTAGGTAAACTTCA 3'
[0476] For each SNP, the first primer contained a recognition site
for the restriction enzyme EcoRI and had a biotin tag at the
extreme 5' end. The second primer used to amplify each SNP
contained a recognition site for the restriction enzyme BceA I.
PCR Reaction
[0477] The PCR reactions were performed as described in Example 2
except that the following annealing temperatures were used: the
annealing temperature for the first cycle of PCR was 37.degree. C.
for 30 seconds, the annealing temperature for the second cycle of
PCR was 57.degree. C. for 30 seconds, and the annealing temperature
for the third cycle of PCR was 64.degree. C. for 30 seconds. All
subsequent cycles had an annealing temperature of 64.degree. C. for
30 seconds. Thirty seven (37) cycles of PCR were performed. After
PCR, 1/4 of the volume was removed from each reaction, and combined
into a single tube.
Purification of Fragment of Interest
[0478] The PCR products (now combined into one sample, and referred
to as "the sample") were separated from the genomic template DNA as
described in Example 2 except that the sample was bound to a single
well of a Streptawell microtiter plate.
Restriction Enzyme Digestion of Isolated Fragments
[0479] The sample was digested with the restriction enzyme BceA I,
which bound the recognition site in the second primer. The
restriction enzyme digestions were performed following the
instructions supplied with the enzyme. After the restriction enzyme
digest, the wells were washed three times with 1.times.PBS.
Incorporation of Nucleotides
[0480] The restriction enzyme digest described above yielded DNA
molecules with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide in the presence of
a DNA polymerase.
[0481] The following components were used for the fill in reaction:
1 .mu.l of fluorescently labeled ddATP; 1 .mu.l of fluorescently
labeled ddTTP; 1 .mu.l of fluorescently labeled ddGTP; 1 .mu.l of
fluorescently labeled ddCTP; 2 .mu.l of 10.times. sequenase buffer,
0.25 .mu.l of Sequenase, and water as needed for a 20 .mu.l
reaction. The fill in reaction was performed at 40.degree. C. for
10 min. All labeling reagents were obtained from Amersham (Thermo
Sequenase Dye Terminator Cycle Sequencing Core Kit (US 79565); the
concentration of the ddNTPS provided in the kit is proprietary and
not published by Amersham). In the presence of fluorescently
labeled ddNTPs, the 3' recessed end was filled in by one base,
which corresponds to the SNP or locus of interest.
[0482] After the incorporation of nucleotide, the Streptawell was
rinsed with 1.times.PBS (100 .mu.l) three times. The "filled in"
DNA fragments were then released from the Streptawell by digestion
with the restriction enzyme EcoRI following the manufacturer's
instructions. Digestion was performed for 1 hour at 37.degree. C.
with shaking at 120 rpm.
Detection of the Locus of Interest
[0483] After release from the streptavidin matrix, 2-3 .mu.l of the
10 .mu.l sample was loaded in a 48 well membrane tray (The Gel
Company, catalog number TAM48-01). The sample in the tray was
absorbed with a 48 Flow Membrane Comb (The Gel Company, catalog
number AM48), and inserted into a 36 cm 5% acrylamide (urea) gel
(BioWhittaker Molecular Applications, Long Ranger Run Gel Packs,
catalog number 50691).
[0484] The sample was electrophoresed into the gel at 3000 volts
for 3 min. The membrane comb was removed, and the gel was run for 3
hours on an ABI 377 Automated Sequencing Machine. The incorporated
nucleotide was detected by fluorescence.
[0485] The primers were designed so that each amplified locus of
interest differed in size. As shown in FIG. 10, each amplified loci
of interest differed by about 5-10 nucleotides, which allowed the
loci of interest to be separated from one another by gel
electrophoresis. Two nucleotides were detected for SNP TSC0087315,
which were guanine and cytosine. These are the two nucleotides
reported to exist at SNP TSC0087315
(http://snp.cshl.org/snpsearch.shtml). The sample comprised
template DNA from 36 individuals and because the DNA molecules that
incorporated a guanine differed in molecular weight from those that
incorporated a cytosine, distinct bands were seen for each
nucleotide.
[0486] Two nucleotides were detected at SNP HC21S00027, which were
guanine and adenosine (FIG. 10). The two nucleotides reported for
this SNP site are guanine and adenosine
(http://snp.cshl.org/snpsearch.shtml). As discussed above, the
sample contained template DNA from thirty-six individuals, and one
would expect both nucleotides to be represented in the sample. The
molecular weight of the DNA fragments that incorporated a guanine
was distinct from the DNA fragments that incorporated an adenosine,
which allowed both nucleotides to be detected.
[0487] The nucleotide cytosine was detected at SNP TSC0214366 (FIG.
10). The two nucleotides reported to exist at this SNP position are
thymidine and cytosine.
[0488] The nucleotide guanine was detected at SNP TSC0413944 (FIG.
10). The two nucleotides reported for this SNP are guanine and
cytosine (http://spp.cshl.org/snpsearch.shtml).
[0489] The nucleotide cytosine was detected at SNP TS00095512 (FIG.
10). The two nucleotides reported for this SNP site are guanine and
cytosine (http://snp.cshl.org/snpsearch.shtml).
[0490] The nucleotide detected at SNP HC21S00131 was guanine. The
two nucleotides reported for this SNP site are guanine and
adenosine (http://snp.cshl.org/snpsearch.shtml).
[0491] As discussed above, the sample was comprised of DNA
templates from thirty-six individuals and one would expect both
nucleotides at the SNP sites to be represented. For SNP TSC0413944,
TSC0095512, TSC0214366 and HC21S00131, one of the two nucleotides
was detected. It is likely that both nucleotides reported for these
SNP sites are present in the sample but that one fluorescent dye
overwhelms the other. The molecular weight of the DNA molecules
that incorporated one nucleotide did not allow efficient separation
of the DNA molecules that incorporated the other nucleotide.
However, the SNPs were readily separated from one another, and for
each SNP, a proper nucleotide was incorporated. The sequences of
multiple loci of interest from multiple chromosomes, which were
treated as a single sample after PCR, were determined.
[0492] A single reaction containing fluorescently labeled ddNTPs
was performed with the sample that contained multiple loci of
interest. Alternatively, four separate fill in reactions can be
performed where each reaction contains one fluorescently labeled
nucleotide (ddATP, ddTTP, ddGTP, or ddCTP) and unlabeled ddNTPs
(see Example 2, FIGS. 7A-7D and FIGS. 9A-C). Four separate "fill
in" reactions will allow detection of any nucleotide that is
present at the loci of interest. For example, if analyzing a sample
that contains multiple loci of interest from a single individual,
and said individual is heterozygous at one or more than one loci of
interest, four separate "fill in" reactions can be used to
determine the nucleotides at the heterozygous loci of interest.
[0493] Also, when analyzing a sample that contains templates from
multiple individuals, four separate "fill in" reactions will allow
detection of nucleotides present in the sample, independent of how
frequent the nucleotide is found at the locus of interest. For
example, if a sample contains DNA templates from 50 individuals,
and 49 of the individuals have a thymidine at the locus of
interest, and one individual has a guanine, the performance of four
separate "fill in" reactions, wherein each "fill in" reaction is
run in a separate lane of a gel, such as in FIGS. 9A-9C, will allow
detection of the guanine. When analyzing a sample comprised of
multiple DNA templates, multiple "fill in" reactions will alleviate
the need to distinguish multiple nucleotides at a single site of
interest by differences in mass.
[0494] In this example, multiple single nucleotide polymorphisms
were analyzed. It is also possible to determine the presence or
absence of mutations, including but not limited to point mutations,
transitions, transversions, translocations, insertions, and
deletions from multiple loci of interest. The multiple loci of
interest can be from a single chromosome or from multiple
chromosomes. The multiple loci of interest can be from a single
gene or from multiple genes.
[0495] The sequence of multiple loci of interest that cause or
predispose to a disease phenotype can be determined. For example,
one could amplify one to tens to hundreds to thousands of genes
implicated in cancer or any other disease. The primers can be
designed so that each amplified loci of interest differs in size.
After PCR, the amplified loci of interest can be combined and
treated as a single sample. Alternatively, the multiple loci of
interest can be amplified in one PCR reaction or the total number
of loci of interest, for example 100, can be divided into samples,
for example 10 loci of interest per PCR reaction, and then later
pooled. As demonstrated herein, the sequence of multiple loci of
interest can be determined. Thus, in one reaction, the sequence of
one to ten to hundreds to thousands of genes that predispose or
cause a disease phenotype can be determined.
Example 4
[0496] The ability to determine the sequence or detect chromosomal
abnormalities of a fetus using free fetal DNA in a sample from a
pregnant female has been hindered by the low percentage of free
fetal DNA. Increasing the percentage of free fetal DNA would
enhance the detection of mutation, insertion, deletion,
translocation, transversion, monosomy, trisomy, trisomy 21, trisomy
18, trisomy 13, XXY, XXX, other aneuploidies, deletion, addition,
amplification, translocation and rearrangement. The percent of
fetal DNA in plasma obtained from a pregnant female was determined
both in the absence and presence of inhibitors of cell lysis. A
genetic marker on the Y chromosome was used to calculate the
percent of fetal DNA.
Preparation of Template DNA
[0497] The DNA template was prepared from a 5 ml sample of blood
obtained by venipuncture from a human volunteer with informed
consent. The blood was aliquoted into two tubes (Fischer
Scientific, 9 ml EDTA Vacuette tubes, catalog number NC9897284).
Formaldehyde (25 .mu.l/ml of blood) was added to one of the tubes.
The sample in the other tube remained untreated, except for the
presence of the EDTA. The tubes were spun at 1000 rpm for ten
minutes. Two milliliters of the supernatant (the plasma) of each
sample was transferred to a new tube and spun at 3000 rpm for ten
minutes. 800 .mu.l of each sample was used for DNA purification.
DNA was isolated using the Qiagen Midi Kit for purification of DNA
from blood cells (QIAmp DNA Blood Midi Kit, Catalog number 51183).
DNA was eluted in 100 .mu.l of distilled water. Two DNA templates
were obtained: one from the blood sample treated with EDTA, and one
from the blood sample treated with EDTA and formaldehyde.
Primer Design
[0498] Two different sets of primers were used: one primer set was
specific for the Y chromosome, and thus specific for fetal DNA, and
the other primer set was designed to amplify the cystic fibrosis
gene, which is present on both maternal template DNA and fetal
template DNA.
[0499] In this example, the first and second primers were designed
so that the entire 5' and 3' sequence of each primer annealed to
the template DNA. In this example, the fetus had an XY genotype,
and the Y chromosome was used as a marker for the presence of fetal
DNA. The following primers were designed to amplify the SRY gene on
the Y chromosome.
TABLE-US-00050 First primer: (SEQ ID NO: 263) 5'
TGGCGATTAAGTCAAATTCGC 3' Second primer: (SEQ ID NO: 264) 5
CCCCCTAGTACCCTGACAATGTATT 3'
[0500] Primers designed to amplify any gene, or region of a region,
or any part of any chromosome could be used to detect maternal and
fetal DNA. In this example, the following primers were designed to
amplify the cystic fibrosis gene:
TABLE-US-00051 First primer: (SEQ ID NO: 265) 5'
CTGTTCTGTGATATTATGTGTGGT 3' Second primer: (SEQ ID NO: 266) 5'
AATTGTTGGCATTCCAGCATTG 3'
PCR Reaction
[0501] The SRY gene and the cystic fibrosis gene were amplified
from the template genomic DNA using PCR (U.S. Pat. Nos. 4,683,195
and 4,683,202). For increased specificity, a "hot-start" PCR was
used. PCR reactions were performed using the HotStarTaq Master Mix
Kit supplied by Qiagen (Catalog No. 203443). For amplification of
the SRY gene, the DNA eluted from the Qiagen purification column
was diluted serially 1:2. For amplification of the cystic fibrosis
gene, the DNA from the Qiagen purification column was diluted 1:4,
and then serially diluted 1:2. The following components were used
for each PCR reaction: 8 .mu.l of template DNA (diluted or
undiluted), 1 .mu.l of each primer (5 .mu.M), 10 .mu.l of HotStar
Tag mix. The following PCR conditions were used:
[0502] (1) 950 C for 15'
[0503] (2) 94.degree. C. for 1'
[0504] (3) 54.degree. C. for 15''
[0505] (4) 72.degree. C. for 30''
[0506] (5) Repeat steps 2-4 for 45 cycles.
[0507] (6) 10' at 72.degree. C.
Quantification of Fetal DNA
[0508] The DNA templates that were eluted from the Qiagen columns
were serially diluted to the following concentrations: 1:2, 1:4,
1:8, 1:16, 1:32, 1:64, 1:128, 1:256, 1:512, 1:1024, 1:2048, and
1:4096. Amplification of the SRY gene was performed using the
templates that were undiluted, 1:2, 1:4, 1:8, 1:16, 1:32, 1:64,
1:128, 1:256, 1:512. Amplification of the cystic fibrosis gene was
performed using the DNA templates that were diluted 1:4, 1:8, 1:16,
1:32, 1:64, 1:128, 1:256, 1:512, 1:1024, 1:2048, and 1:4096. The
same dilution series was performed with the DNA templates that were
purified from the plasma sample treated with EDTA alone and the
plasma sample treated with EDTA and formaldehyde.
[0509] The results of the PCR reactions using the DNA template that
was isolated from the plasma sample treated with EDTA are shown in
FIG. 11A. The SRY gene was amplified from the undiluted DNA
template, and also in the sample that was diluted 1:2 (FIG. 11A).
The SRY gene was not amplified in the next seven serial dilutions.
On the other hand, the cystic fibrosis gene was detected in the
serial dilutions up to 1:256. A greater presence of the cystic
fibrosis gene was expected because of the higher percentage of
maternal DNA present in the plasma. The last dilution sample that
provided for amplification of the gene product was assumed to have
one copy of the cystic fibrosis gene or the SRY gene.
[0510] The results of the PCR reactions using the DNA template that
was isolated from the plasma sample treated with formaldehyde and
EDTA are shown in FIG. 11B. The SRY gene was amplified from the
undiluted DNA template, and also in the sample that was diluted 1:2
(FIG. 11B). The SRY gene was not amplified in the next six
dilutions. However, in the 1:256 dilution, the SRY gene was
detected. It is unlikely that the amplification in the 1:256 sample
represents a real signal because the prior six dilution series were
all negative for amplification of SRY. Amplification of the SRY
gene in this sample was likely an experimental artifact resulting
from the high number of PCR cycles used. Thus, the 1:256 sample was
not used in calculating the amount of fetal DNA present in the
sample.
[0511] Amplification of the cystic fibrosis gene was detected in
the sample that was diluted 1:16 (FIG. 11B). The presence of the
formal in prevents maternal cell lysis, and thus, there is a lower
percentage of maternal DNA in the sample. This is in strong
contrast to the sample that was treated with only EDTA, which
supported amplification up to a dilution of 1:256.
[0512] The percent of fetal DNA present in the maternal plasma was
calculated using the following formula:
% fetal DNA=(amount of SRY gene/amount of cystic fibrosis
gene)*2*100.
The amount of SRY gene was represented by the highest dilution
value in which the gene was amplified. Likewise, the amount of
cystic fibrosis gene was represented by the highest dilution value
in which it was amplified. The formula contains a multiplication
factor of two (2), which is used to normalize for the fact that
there is only one copy of the SRY gene (located on the Y
chromosome), while there are two copies of the cystic fibrosis
gene.
[0513] For the above example, the percentage of fetal DNA present
in the sample that was treated with only EDTA was 1.56%
(2/256*2*100). The reported percentage of fetal DNA present in the
plasma is between 0.39-11.9% (Pertl and Bianchi, Obstetrics and
Gynecology, Vol, 98, No. 3, 483-490 (2001). The percentage of fetal
DNA present in the sample treated with formalin and EDTA was 25%
(2/16*2*100). The experiment was repeated numerous times, and each
time the presence of formalin increased the overall percentage of
fetal DNA.
[0514] The percent fetal DNA from eighteen blood samples with and
without formalin was calculated as described above with the
exception that serial dilutions of 1:5 were performed. As 1:5
dilutions were performed, the last serial dilution that allowed
detection of either the SRY gene or the cystic fibrosis gene may
have had one copy of the gene or it may have had 4 copies of the
gene. The results from the eighteen samples with and without
formalin are summarized in Table V. The low range assumes that the
last dilution sample had one copy of the genes and the high range
assumes that the last dilution had four copies of the genes.
TABLE-US-00052 TABLE V Mean Percentage Fetal DNA with and without
formalin. Sample Lower Range Upper Range Formalin 19.47 43.69
Without Formalin 7.71 22.1
[0515] An overall increase in fetal DNA was achieved by reducing
the maternal cell lysis, and thus, reducing the amount of maternal
DNA present in the sample. In this example, formaldehyde was used
to prevent lysis of the cells, however any agent that prevents the
lysis of cells or increases the structural integrity of the cells
can be used. Two or more than two cell lysis inhibitors can be
used. The increase in fetal DNA in the maternal plasma allows the
sequence of the fetal DNA to be determined, and provides for the
rapid detection of abnormal DNA sequences or chromosomal
abnormalities including but not limited to point mutation, reading
frame shift, transition, transversion, addition, insertion,
deletion, addition-deletion, frame-shift, missense, reverse
mutation, and microsatellite alteration, trisomy, monosomy, other
aneuploidies, amplification, rearrangement, translocation,
transversion, deletion, addition, amplification, fragment,
translocation, and rearrangement.
Example 5
[0516] A DNA template from an individual with a genotype of trisomy
21 was analyzed. Three loci of interest were analyzed on chromosome
13 and two loci of interest were analyzed on chromosome 21.
Preparation of Template DNA
[0517] The template DNA was prepared from a 5 ml sample of blood
obtained by venipuncture from a human volunteer with informed
consent. The human volunteer had previously been genotyped to have
an additional chromosome 21 (trisomy 21). Template DNA was isolated
using QIAamp DNA Blood Midi Kit supplied by QIAGEN (Catalog number
51183).
Primer Design
[0518] The following five single nucleotide polymorphisms were
analyzed: SNP TSC 0115603 located on chromosome 21; SNP TSC
03209610 located on chromosome 21; SNP TSC 0198557 located on
chromosome 13; and SNP TSC 0200347 located on chromosome 13. The
DNA template from another individual was used as an internal
control. The SNP TSC 0200347, which was previously identified as
being homozygous for guanine, was used as the internal control. The
SNP Consortium Ltd database can be accessed at
http://snp.cshl.org/, website address effective as of Apr. 1,
2002.
[0519] SNP TSC 0115603 was amplified using the following
primers:
TABLE-US-00053 First Primer: (SEQ ID NO: 267) 5'
GTGCACTTACGTGAATTCAGATGAACGTGATGTAGTAG 3' Second Primer: (SEQ ID
NO: 268) 5' TCCTCGTACTCAACGGCTTTCTCTGAAT 3'
[0520] The first primer was biotinylated at the 5' end, and
contained the restriction enzyme recognition site for EcoR I. The
second primer contained the restriction enzyme recognition site for
the restriction enzyme BceA I.
[0521] SNP TSC 0309610 was amplified using the following
primers:
TABLE-US-00054 First primer: (SEQ ID NO: 269) 5'
TCCGGAACACTAGAATTCTTATTTACATACACACTTGT 3' Second primer: (SEQ ID
NO: 270) 5' CGAATAAGGTAGACGGCAACAATGAGAA 3'
[0522] The first primer contained a biotin group at the 5' end, and
a restriction enzyme recognition site for the restriction enzyme
EcoR I. The second primer contained the restriction enzyme
recognition site for BceA I.
[0523] Submitted SNP (ss) 813773 (accession number assigned by the
NCBI Submitted SNP (ss) Database) was amplified with the following
primers:
TABLE-US-00055 First primer: (SEQ ID NO: 271) 5'
CGGTAAATCGGAGAATTCAGAGGATTTAGAGGAGCTAA 3' Second primer: (SEQ ID
NO: 272) 5' CTCACGTTCGTTACGGCCATTGTGATAGC 3'
[0524] The first primer contains a biotin group at the 5' end, and
a recognition site for the restriction enzyme EcoR I. The second
primer contained the restriction enzyme recognition site for BceA
I.
[0525] SNP TSC 0198557 was amplified with the following
primers:
TABLE-US-00056 First primer: (SEQ ID NO: 273) 5'
GGGGAAACAGTAGAATTCCATATGGACAGAGCTGTACT 3' Second primer: (SEQ ID
NO: 274) 5' TGAAGCTGTCGGACGGCCTTTGCCCTCTC 3'
[0526] The first primer contains a biotin group at the 5' end, and
a recognition site for the restriction enzyme EcoR I. The second
primer contained the restriction enzyme recognition site for BceA
I.
[0527] SNP TSC 0197279 was amplified with the following
primers:
TABLE-US-00057 First primer: (SEQ ID NO: 275) 5'
ATGGGCAGTTATGAATTCACTACTCCCTGTAGCTTGTT 3' Second primer: (SEQ ID
NO: 276) 5' TGATTGGCGCGAACGGCACTCAGAGAAGA 3'
[0528] The first primer contained a biotin group at the 5' end, and
a recognition site for the restriction enzyme for EcoR I. The
second primer contained the recognition site for the restriction
enzyme BceA I.
[0529] SNP TSC 0200347 was amplified with the following
primers:
TABLE-US-00058 First primer: (SEQ ID NO: 277) 5'
CTCAAGGGGACCGAATTCGCTGGGGTCTTCTGTGGGTC 3' Second primer: (SEQ ID
NO: 278) 5' TAGGGCGGCGTGACGGCCAGCCAGTGGT 3'
[0530] The first primer contained a biotin group at the 5' end, and
the recognition site for the restriction enzyme EcoR I. The second
primer contained the restriction enzyme recognition site for BceA
I.
PCR Reaction
[0531] All five loci of interest were amplified from the template
genomic DNA using PCR (U.S. Pat. Nos. 4,683,195 and 4,683,202). For
increased specificity, a "hot-start" PCR was used. PCR reactions
were performed using the HotStarTaq Master Mix Kit supplied by
QIAGEN (catalog number 203443). The amount of template DNA and
primer per reaction can be optimized for each locus of interest; in
this example, 40 ng of template human genomic DNA and 5 .mu.M of
each primer were used. Thirty-eight cycles of PCR were performed.
The following PCR conditions were used for SNP TSC 0115603, SNP TSC
0309610, and SNP TSC 02003437:
[0532] (1) 95.degree. C. for 15 minutes and 15 seconds;
[0533] (2) 42.degree. C. for 30 seconds;
[0534] (3) 95.degree. C. for 30 seconds;
[0535] (4) 60.degree. C. for 30 seconds;
[0536] (5) 95.degree. C. for 30 seconds;
[0537] (6) 69.degree. C. for 30 seconds;
[0538] (7) 95.degree. C. for 30 seconds;
[0539] (8) Repeat steps 6 and 7 thirty nine (37) times;
[0540] (9) 72.degree. C. for 5 minutes.
[0541] The following PCR conditions were used for SNP ss813773, SNP
TSC 0198557, and SNP TSC 0197279:
[0542] (1) 95.degree. C. for 15 minutes and 15 seconds;
[0543] (2) 37.degree. C. for 30 seconds;
[0544] (3) 95.degree. C. for 30 seconds;
[0545] (4) 57.degree. C. for 30 seconds;
[0546] (5) 95.degree. C. for 30 seconds;
[0547] (6) 64.degree. C. for 30 seconds;
[0548] (7) 95.degree. C. for 30 seconds;
[0549] (8) Repeat steps 6 and 7 thirty nine (37) times; and
[0550] (9) 72.degree. C. for 5 minutes.
[0551] In the first cycle of each PCR, the annealing temperature
was about the melting temperature of the 3' annealing region of the
second primer. The annealing temperature in the second cycle of PCR
was about the melting temperature of the 3' region, which anneals
to the template DNA, of the first primer. The annealing temperature
in the third cycle of PCR was about the melting temperature of the
entire sequence of the second primer. Escalating the annealing
temperature from TM1 to TM2 to TM3 in the first three cycles of PCR
greatly improves specificity. These annealing temperatures are
representative, and the skilled artisan will understand the
annealing temperatures for each cycle are dependent on the specific
primers used. The temperatures and times for denaturing, annealing,
and extension, can be optimized by trying various settings and
using the parameters that yield the best results.
Purification of Fragment of Interest
[0552] PCR products were separated from the components of the PCR
reaction using Qiagen's MinElute PCR Purification Kit following
manufacturer's instructions (Catalog number 28006). The PCR
products were eluted in 20 .mu.l of distilled water. For each
amplified SNP, one microliter of PCR product, 1 .mu.l of amplified
internal control DNA (SNP TSC 0200347), and 8 .mu.l of distilled
water were mixed. Five microliters of each sample was placed into
two separate reaction wells of a Pierce StreptaWell Microtiter
plate (catalog number 15501). The first primers contained a 5'
biotin tag so the PCR products bound to the Streptavidin coated
wells while the genomic template DNA did not. The streptavidin
binding reaction was performed using a Thermomixer (Eppendorf) at
150 rpm for 1 hour at 45.degree. C. Each well was aspirated to
remove unbound material, and washed three times with 1.times.PBS,
with gentle mixing (Kandpal et al., Nucl. Acids Res. 18:1789-1795
(1990); Kaneoka et al., Biotechniques 10:30-34 (1991); Green et
al., Nucl. Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[0553] The purified PCR products were digested with the restriction
enzyme that bound the recognition site that was incorporated into
the PCR products from the second primer. The purified PCR products
were digested with the restriction enzyme BceA I (New England
Biolabs, catalog number R0623S). The digests were performed in the
wells of the microtiter plate following the instructions supplied
with the restriction enzyme. After digestion with the appropriate
restriction enzyme, the wells were washed three times with PBS to
remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[0554] The restriction enzyme digest described above yielded a DNA
fragment with a 5' overhang, which contained the SNP and a 3'
recessed end. The 5' overhang functioned as a template allowing
incorporation of a nucleotide or nucleotides in the presence of a
DNA polymerase.
[0555] For each SNP, two fill in reactions were performed; each
reaction contained a different fluorescently labeled
dideoxynucleotide (ddATP, ddCTP, ddGTP, or ddTTP, depending on the
reported nucleotides to exist at a particular SNP). For example,
the nucleotides adenine and thymidine have been reported at SNP TSC
0115603. Therefore, the digested PCR product for SNP TSC 0115603
was mixed with either fluorescently labeled ddATP or fluorescently
labeled ddTTP. Each reaction contained fluorescently labeled ddGTP
for the internal control. The following components were added to
each fill in reaction: 2 .mu.l of a ROX-conjugated
dideoxynucleotide (depending on the nucleotides reported for each
SNP), 2 .mu.l of ROX-conjugated ddGTP (internal control), 2.5 .mu.l
of 10.times. sequenase buffer, 2 .mu.l of Sequenase, and water as
needed for a 25 .mu.l reaction. All of the fill in reactions were
performed at 45.degree. C. for 45 min. However, shorter time
periods of incorporation can be used. Non-fluorescently labeled
ddNTPs were purchased from Fermentas Inc. (Hanover, Md.). The
ROX-conjugated ddNTPs were obtained from Perkin Elmer. In the
presence of fluorescently labeled ddNTPs, the 3' recessed end was
extended by one base, which corresponds to the SNP or locus of
interest.
[0556] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoR I following manufacturer's recommendations. Digestion
was performed for 1 hour at 37.degree. C. with shaking at 120
rpm.
Detection of the Locus of Interest
[0557] After release from the streptavidin matrix, 3 .mu.l of the
10 .mu.l sample was loaded in a 48 well membrane tray (The Gel
Company, catalog number TAM48-01). The sample in the tray was
absorbed with a 48 Flow Membrane Comb (The Gel Company, catalog
number AM48), and inserted into a 36 cm 5% acrylamide (urea) gel
(BioWhittaker Molecular Applications, Long Ranger Run Gel Packs,
catalog number 50691).
[0558] The sample was electrophoresed into the gel at 3000 volts
for 3 min. The membrane comb was removed, and the gel was run for 3
hours on an ABI 377 Automated Sequencing Machine. The incorporated
labeled nucleotide was detected by fluorescence.
[0559] As seen in FIG. 12, SNP TSC 0115603 was "filled in" with
labeled ddTTP (lane 1) and in a separate reaction with labeled
ddATP (lane 3). The calculated ratio between the nucleotides, using
the raw data, was 66:34, which is consistent with the theoretical
ratio of 66:33 for a SNP on chromosome 21 in an individual with
trisomy 21. Both the ddTTP and ddATP were labeled with the same
fluorescent dye to minimize variability in incorporation
efficiencies of the dyes. However, nucleotides with different
fluorescent labels or any detectable label can be used. It is
preferable to calculate the coefficients of incorporation when
different labels are used.
[0560] Each fill in reaction was performed in a separate well so it
was possible that there could be variability in DNA binding between
the wells of the microtiter plate. To account for the potential
variability of DNA binding to the streptavidin-coated plates, an
internal control was used. The internal control (SNP TSC 0200347),
which is homozygous for guanine, was added to the sample prior to
splitting the sample into two separate wells, and thus, an equal
amount of the internal control should be present in each well. The
amount of incorporated ddGTP can be fixed between the two
reactions. If the amount of DNA in each well is equal, the amount
of incorporated ddGTP should be equal because the reaction is
performed under saturating conditions, with saturating conditions
being defined as conditions that support incorporation of a
nucleotide at each template molecule. Using the internal control,
the ratio of incorporated ddATP to ddTTP was 614:36.6, This ratio
was very similar to the ratio obtained with the raw data,
indicating that there are minor differences in the two fill in
reactions for a particular SNP.
TABLE-US-00059 TABLE VI Allele Frequencies at Multiple SNPs on DNA
Template from Individual with Trisomy 21 Peak Allele Internal
Allele SNP Allele Area Ratio Control Normalized Peak Area Ratio (%)
TSC A 5599 66 723 5599 63.4 0115603 T 2951 34 661 3227
((723/661)*2951) 36.6 TSC T 4126 64 1424 4126 66.8 0309610 C 2342
36 1631 2045 ((1424/1631)*2342) 33.2 ss813773 A 4199 46 808 4199 41
C 4870 54 647 6082 ((808/647)*4870) 59 TSC T 3385 55 719 3385 49
0198557 C 2741 45 559 3525 719/559 *2741) 51 TSC T 8085 53 2752
8085 50.7 0197279 C 7202 47 2520 7865 (2752/2520 *7202 49.3
[0561] SNP TSC 0309610 was filled in with ddTTP (lane 3) or ddCTP
(lane 4) (FIG. 12). The calculated ratio for the nucleotides, using
the raw data, was 64:36. Both ddTTP and ddCTP were labeled with the
same fluorescent dye. After normalization to the internal control,
as discussed above, the calculated allele ratio of ddTTP to ddCTP
was 66.8:33.2 (Table VI). Again, both the calculated ratio from the
raw data and the calculated ratio using the internal control are
very similar to the theoretical ratio of 66.6:33.4 for a SNP on
chromosome 21 in an individual with trisomy.
[0562] To demonstrate that the 66:33 ratios for nucleotides at
heterozygous SNPS represented loci on chromosomes present in three
copies, SNPs on chromosome 13 were analyzed. The individual from
whom the blood sample was obtained had previously been genotyped
with one maternal chromosome 13, and one paternal chromosome
13.
[0563] Submitted SNP (ss) 813773 was filled in with ddATP (lane 5)
or ddCTP (lane 6) (FIG. 12). The calculated ratio for the
nucleotides at this heterozygous SNP, using the raw data, was
46:54. This ratio is within 10% of the expected ratio of 50:50.
Importantly, the ratio does not approach the 66:33 ratio expected
when there is an additional copy of a chromosome.
[0564] After normalization to the internal control, the calculated
ratio was 41:59. Contrary to the expected result, normalization to
the internal control increased the discrepancy between the
calculated ratio and the theoretical ratio. This result may
represent experimental error that occurred in aliquoting the DNA
samples.
[0565] Also, it is possible that the restriction enzyme used to
generate the overhang, which was used as a template for the
"fill-in" reaction, preferentially cut one DNA template over the
other DNA template. The two templates differ, with respect to the
nucleotide at the SNP site, and this may influence the cutting. The
primers can be designed such that the nucleotides adjacent to the
cut site are the same, independent of the nucleotide at the SNP
site (discussed further in the section entitled "Primer
Design").
[0566] SNP TSC 0198557, which is on chromosome 13, was filled in
with ddTTP (lane 7) in one reaction and ddCTP (lane 8) in another
(FIG. 12). The calculated ratio for the nucleotides at this SNP,
using the raw data, was 55:45. After normalization to the internal
control, the calculated allele ratio of T:C was 49:51. The
normalized ratio was closer to the theoretical ratio of 50:50 for
an individual with two copies of chromosome 13.
[0567] SNP TSC 0197279, which is on chromosome 13, was filled in
with ddTTP (lane 9) in one reaction and ddCTP (lane 10) in another
(FIG. 12). The calculated ratio for the nucleotides at this SNP,
using the raw data was 53:47. After normalization to the internal
control, the calculated allele ratio of T:C was 50.7:49.3. This is
consistent with the theoretical ratio of 50:50 for an individual
with only two copies of chromosome 13.
[0568] The ratio for the nucleotides at two of the analyzed SNPs on
chromosome 13 was approximately 50:50. One SNP, ss813773, showed a
ratio of 46:54, and when normalized to the internal control, the
ratio was 41:59. These ratios deviate from the expected 50:50, but
at the same time, the ratios are not indicative of an extra
chromosome, which is indicated with a ratio of 66:33. While the
data from this particular SNP is inconclusive, it does not
represent a false positive. No conclusion could be drawn on the
data from this SNP. However, the other two SNPs provided data that
indicated a normal number of chromosomes. It is preferable to
analyze multiple SNPs on a chromosome including but not limited to
1-5, 5-10, 10-50, 50-100, 100-200, 200-300, 300-400, 400-500,
500-600, 600-700, 700-800, 800-900, 900-1000, 1000-2000, 2000-3000,
and greater than 3000. Preferably, the average of the ratios for a
particular chromosome will be used to determine the presence or
absence of a chromosomal abnormality. However, it is still possible
to analyze one locus of interest. In the event that inconclusive
data is obtained, another locus of interest can be analyzed.
[0569] The individual from whom the DNA template was obtained had
previously been genotyped with trisomy 21, and the allele
frequencies at SNPs on chromosome 21 indicate the presence of an
additional chromosome 21. The additional chromosome contributes an
additional nucleotide for each SNP, and thus alters the traditional
50:50 ratio at a heterozygous SNP. These results are consistent for
multiple SNPs, and are specific for those found on chromosome 21.
The allele frequencies for SNPs on chromosome 13 gave the expected
ratios of approximately 50:50. These results demonstrate that this
method of SNP detection can be used to detect chromosomal
abnormalities including but not limited to translocations,
transversions, monosomies, trisomy 21, trisomy 18, trisomy 13,
other aneuploidies, deletions, additions, amplifications,
translocations and rearrangements.
Example 6
[0570] Genomic DNA was obtained from four individuals after
informed consent was obtained. Six SNPs on chromosome 13
(TSC0837969, TSC0034767, TSC 1130902, TSC0597888, TSC0195492,
TSC0607185) were analyzed using the template DNA. Information
regarding these SNPs can be found at the following website
www.snp.chsl.org/snpsearch.shtml; website active as of Feb. 11,
2003).
[0571] A single nucleotide labeled with one fluorescent dye was
used to genotype the individuals at the six selected SNP sites. The
primers were designed to allow the six SNPs to be analyzed in a
single reaction.
Preparation of Template DNA
[0572] The template DNA was prepared from a 9 ml sample of blood
obtained by venipuncture from a human volunteer with informed
consent. Template DNA was isolated using the QIAmp DNA Blood Midi
Kit supplied by QIAGEN (Catalog number 51183). The template DNA was
isolated as per instructions included in the kit.
Design of Primers
[0573] SNP TSC0837969 was amplified using the following primer
set:
TABLE-US-00060 First primer: (SEQ ID NO: 30) 5'
GGGCTAGTCTCCGAATTCCACCTATCCTACCAAATGTC 3' Second primer: (SEQ ID
NO: 31) 5' TAGCTGTAGTTAGGGACTGTTCTGAGCAC 3'
[0574] The first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI. The
first primer was designed to anneal 44 bases from the locus of
interest. The second primer contained a restriction enzyme
recognition site for BsmF I.
[0575] SNP TSC0034767 was amplified using the following primer
set:
TABLE-US-00061 First primer: (SEQ ID NO: 32) 5'
CGAATGCAAGGCGAATTCGTTAGTAATAACACAGTGCA 3' Second primer: (SEQ ID
NO: 33) 5' AAGACTGGATCCGGGACCATGTAGAATAC 3'
[0576] The first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI. The
first primer was designed to anneal 50 bases from the locus of
interest. The second primer contained a restriction enzyme
recognition site for BsmF I.
[0577] SNP TSC1130902 was amplified using the following primer
set:
TABLE-US-00062 First primer: (SEQ ID NO: 34) 5'
TCTAACCATTGCGAATTCAGGGCAAGGGGGGTGAGATC 3' Second primer: (SEQ ID
NO: 35) 5' TGACTTGGATCCGGGACAACGACTCATCC 3'
[0578] The first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI. The
first primer was designed to anneal 60 bases from the locus of
interest. The second primer contained a restriction enzyme
recognition site for BsmF I.
[0579] SNP TSC0597888 was amplified using the following primer
set:
TABLE-US-00063 First primer: (SEQ ID NO: 36) 5'
ACCCAGGCGCCAGAATTCTTTAGATAAAGCTGAAGGGA 3' Second primer: (SEQ ID
NO: 37) 5' GTTACGGGATCCGGGACTCCATATTGATC 3'
[0580] The first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI. The
first primer was designed to anneal 70 bases from the locus of
interest. The second primer contained a restriction enzyme
recognition site for BsmF I.
[0581] SNP TSC0195492 was amplified using the following primer
set:
TABLE-US-00064 First primer: (SEQ ID NO: 38)
5'CGTTGGCTTGAGGAATTCGACCAAAAGAGCCAAGAGAA Second primer: (SEQ ID NO:
39) 5' AAAAAGGGATCCGGGACCTTGACTAGGAC 3'
[0582] The first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI. The
first primer was designed to anneal 80 bases from the locus of
interest. The second primer contained a restriction enzyme
recognition site for BsmF I.
[0583] SNP TSC0607185 was amplified using the following primer
set:
TABLE-US-00065 First primer: (SEQ ID NO: 40) 5'
ACTTGATTCCGTGAATTCGTTATCAATAAATCTTACAT 3' Second primer: (SEQ ID
NO: 41) 5' CAAGTTGGATCCGGGACCCAGGGCTAACC 3'
[0584] The first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI. The
first primer was designed to anneal 90 bases from the locus of
interest. The second primer contained a restriction enzyme
recognition site for BsmF I.
[0585] All loci of interest were amplified from the template
genomic DNA using the polymerase chain reaction (PCR, U.S. Pat.
Nos. 4,683,195 and 4,683,202, incorporated herein by reference). In
this example, the loci of interest were amplified in separate
reaction tubes but they could also be amplified together in a
single PCR reaction. For increased specificity, a "hot-start" PCR
was used. PCR reactions were performed using the HotStarTaq Master
Mix Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction can be optimized for each
locus of interest but in this example, 40 ng of template human
genomic DNA and 5 .mu.M of each primer were used. Forty cycles of
PCR were performed. The following PCR conditions were used: [0586]
(1) 95.degree. C. for 15 minutes and 15 seconds; [0587] (2)
37.degree. C. for 30 seconds; [0588] (3) 95.degree. C. for 30
seconds; [0589] (4) 57.degree. C. for 30 seconds; [0590] (5)
95.degree. C. for 30 seconds; [0591] (6) 64.degree. C. for 30
seconds; [0592] (7) 95.degree. C. for 30 seconds; [0593] (8) Repeat
steps 6 and 7 thirty nine (39) times; [0594] (9) 72.degree. C. for
5 minutes.
[0595] In the first cycle of PCR, the annealing temperature was
about the melting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[0596] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results. In this example, the
first primer was designed to anneal at various distances from the
locus of interest. The skilled artisan understands that the
annealing location of the first primer can be 5-10, 11-15, 16-20,
21-25, 26-30, 31-35, 36-40, 41-45, 46-50, 51-55, 56-60, 61-65,
66-70, 71-75, 76-80, 81-85, 86-90, 91-95, 96-100, 101-105, 106-110,
111-115, 116-120, 121-125, 126-130, 131-140, 1410-160, 1610-180,
1810-200, 2010-220, 2210-240, 2410-260, 2610-280, 2810-300,
3010-350, 3510-400, 4010-450, 450-500, or greater than 500 bases
from the locus of interest.
Purification of Fragment of Interest
[0597] The PCR products were separated from the genomic template
DNA. After the PCR reaction, 1/4 of the volume of each PCR reaction
from one individual was mixed together in a well of a Streptawell,
transparent, High-Bind plate from Roche Diagnostics GmbH (catalog
number 1 645 692, as listed in Roche Molecular Biochemicals, 2001
Biochemicals Catalog). The first primers contained a 5' biotin tag
so the PCR products bound to the Streptavidin coated wells while
the genomic template DNA did not. The streptavidin binding reaction
was performed using a Thermomixer (Eppendorf) at 1000 rpm for 20
min. at 37.degree. C. Each well was aspirated to remove unbound
material, and washed three times with 1.times.PBS, with gentle
mixing (Kandpal et al., Nucl. Acids Res. 18:1789-1795 (1990);
Kaneoka et al., Biotechniques 10:30-34 (1991); Green et al, Nucl.
Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[0598] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in the Streptawells following the instructions supplied
with the restriction enzyme. After digestion, the wells were washed
three times with PBS to remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[0599] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[0600] Below, a schematic of the 5' overhang for SNP TSC0837969 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang (where R indicates the variable
site).
TABLE-US-00066 5' TTAA 3' AATT R A C A Overhang position 1 2 3
4
[0601] The observed nucleotides for TSC0837969 on the 5' sense
strand (here depicted as the top strand) are adenine and guanine.
The third position in the overhang on the antisense strand
corresponds to cytosine, which is complementary to guanine. As this
variable site can be adenine or guanine, fluorescently labeled
ddGTP in the presence of unlabeled dCTP, dTTP, and dATP was used to
determine the sequence of both alleles. The fill-in reactions for
an individual homozygous for guanine, homozygous for adenine or
heterozygous are diagrammed below.
[0602] Homozygous for Guanine at TSC 0837969:
TABLE-US-00067 Allele 1 5' TTAA G* 3' AATT C A C A Overhang
position 1 2 3 4 Allele 2 5' TTAA G* 3' AATT C A C A Overhang
position 1 2 3 4
[0603] Labeled ddGTP is incorporated into the first position of the
overhang. Only one signal is seen, which corresponds to the
molecules filled in with labeled ddGTP at the first position of the
overhang.
[0604] Homozygous for Adenine at TSC 0837969:
TABLE-US-00068 Allele 1 5' TTAA A T G* 3' AATT T A C A Overhang
position 1 2 3 4 Allele 2 5' TTAA A T G* 3' AATT T A C A Overhang
position 1 2 3 4
[0605] Unlabeled dATP is incorporated at position one of the
overhang, and unlabeled dTTP is incorporated at position two of the
overhang. Labeled ddGTP was incorporated at position three of the
overhang. Only one signal will be seen; the molecules filled in
with ddGTP at position 3 will have a different molecular weight
from molecules filled in at position one, which allows easy
identification of individuals homozygous for adenine or
guanine.
[0606] Heterozygous at TSC0837969:
TABLE-US-00069 Allele 1 5' TTAA G* 3' AATT C A C A Overhang
position 1 2 3 4 Allele 2 5' TTAA A T G* 3' AATT T A C A Overhang
position 1 2 3 4
[0607] Two signals will be seen; one signal corresponds to the DNA
molecules filled in with ddGTP at position 1, and a second signal
corresponding to molecules filled in at position 3 of the overhang.
The two signals can be separated using any technique that separates
based on molecular weight including but not limited to gel
electrophoresis.
[0608] Below, a schematic of the 5' overhang for SNP TSC0034767 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang (where R indicates the variable
site).
TABLE-US-00070 A C A R GTGT 3' CACA 5' 4 3 2 1 Overhang
Position
[0609] The observed nucleotides for TSC0034767 on the 5' sense
strand (here depicted as the top strand) are cytosine and guanine.
The second position in the overhang corresponds to adenine, which
is complementary to thymidine. The third position in the overhang
corresponds to cytosine, which is complementary to guanine.
Fluorescently labeled ddGTP in the presence of unlabeled dCTP,
dTTP, and dATP is used to determine the sequence of both
alleles.
[0610] In this case, the second primer anneals upstream of the
locus of interest, and thus the fill-in reaction occurs on the
anti-sense strand (here depicted as the bottom strand). Either the
sense strand or the antisense strand can be filled in depending on
whether the second primer, which contains the type IIS restriction
enzyme recognition site, anneals upstream or downstream of the
locus of interest.
[0611] Below, a schematic of the 5' overhang for SNP TSC1130902 is
shown. The entire DNA sequence is not reproduced, only a portion to
demonstrate the overhang (where R indicates the variable site).
TABLE-US-00071 5' TTCAT 3' AAGTA R T C C Overhang position 1 2 3
4
[0612] The observed nucleotides for TSC1130902 on the 5' sense
strand (here depicted as the top strand) are adenine and guanine.
The second position in the overhang corresponds to a thymidine, and
the third position in the overhang corresponds to cytosine, which
is complementary to guanine.
[0613] Fluorescently labeled ddGTP in the presence of unlabeled
dCTP, dTTP, and dATP is used to determine the sequence of both
alleles.
[0614] Below, a schematic of the 5' overhang for SNP TSC0597888 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang (where R indicates the variable
site).
TABLE-US-00072 T C T R ATTC 3' TAAG 5' 4 3 2 1 Overhang
position
[0615] The observed nucleotides for TSC0597888 on the 5' sense
strand (here depicted as the top strand) are cytosine and guanine.
The third position in the overhang corresponds to cytosine, which
is complementary to guanine. Fluorescently labeled ddGTP in the
presence of unlabeled dCTP, dTTP, and dATP is used to determine the
sequence of both alleles.
[0616] Below, a schematic of the 5' overhang for SNP TSC0607185 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang (where R indicates the variable
site).
TABLE-US-00073 C C T R TGTC 3' ACAG 5' 4 3 2 1 Overhang
position
[0617] The observed nucleotides for TSC0607185 on the 5' sense
strand (here depicted as the top strand) are cytosine and
thymidine. In this case, the second primer anneals from the locus
of interest, which allows the anti-sense strand to be filled in.
The anti-sense strand (here depicted as the bottom strand) will be
filled in with guanine or adenine.
[0618] The second position in the 5' overhang is thymidine, which
is complementary to adenine, and the third position in the overhang
corresponds to cytosine, which is complementary to guanine.
Fluorescently labeled ddGTP in the presence of unlabeled dCTP,
dTTP, and dATP is used to determine the sequence of both
alleles.
[0619] Below, a schematic of the 5' overhang for SNP TSC0195492 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang.
TABLE-US-00074 5' ATCT 3' TAGA R A C A Overhang position 1 2 3
4
[0620] The observed nucleotides at this site are cytosine and
guanine (here depicted as the top strand). The second position in
the 5' overhang is adenine, which is complementary to thymidine,
and the third position in the overhang corresponds to cytosine,
which is complementary to guanine. Fluorescently labeled ddGTP in
the presence of unlabeled dCTP, dTTP, and dATP is used to determine
the sequence of both alleles.
[0621] As demonstrated above, the sequence of both alleles of the
six SNPs can be determined by labeling with ddGTP in the presence
of unlabeled dATP, dTTP, and dCTP. The following components were
added to each fill in reaction: 1 .mu.l of fluorescently labeled
ddGTP, 0.5 .mu.l of unlabeled dNTPs (40 .mu.M), which contained all
nucleotides except guanine, 2 .mu.l of 10.times. sequenase buffer,
0.25 .mu.l of Sequenase, and water as needed for a 20 .mu.l
reaction. The fill in reaction was performed at 40.degree. C. for
10 min. Non-fluorescently labeled dNTP was purchased from Fermentas
Inc. (Hanover, Md.). All other labeling reagents were obtained from
Amersham (Thermo Sequenase Dye Terminator Cycle Sequencing Core
Kit, US 79565).
[0622] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme. Digestion was performed for 1 hour
at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[0623] After release from the streptavidin matrix, the sample was
loaded into a lane of a 36 cm 5% acrylamide (urea) gel
(BioWhittaker Molecular Applications, Long Ranger Run Gel Packs,
catalog number 50691). The sample was electrophoresed into the gel
at 3000 volts for 3 min. The gel was run for 3 hours on a
sequencing apparatus (Hoefer SQ3 Sequencer). The gel was removed
from the apparatus and scanned on the Typhoon 9400 Variable Mode
Imager. The incorporated labeled nucleotide was detected by
fluorescence.
[0624] As shown in FIG. 11, the template DNA in lanes 1 and 2 for
SNP TSC0837969 is homozygous for adenine. The following fill-in
reaction was expected to occur if the individual was homozygous for
adenine:
[0625] Homozygous for Adenine at TSC 0837969:
TABLE-US-00075 5' TTAA A T G* 3' AATT T A C A Overhang position 1 2
3 4
[0626] Unlabeled dATP was incorporated in the first position
complementary to the overhang. Unlabeled dTTP was incorporated in
the second position complementary to the overhang. Labeled ddGTP
was incorporated in the third position complementary to the
overhang. Only one band was seen, which migrated at about position
46 of the acrylamide gel. This indicated that adenine was the
nucleotide filled in at position one. If the nucleotide guanine had
been filled in, a band would be expected at position 44.
[0627] However, the template DNA in lanes 3 and 4 for SNP
TSC0837969 was heterozygous. The following fill-in reactions were
expected if the individual was heterozygous:
[0628] Heterozygous at TSC0837969:
TABLE-US-00076 Allele 1 5' TTAA G* 3' AATT C A C A Overhang
position 1 2 3 4 Allele 2 5' TTAA A T G* 3' AATT T A C A Overhang
position 1 2 3 4
[0629] Two distinct bands were seen; one band corresponds to the
molecules filled in with ddGTP at position 1 complementary to the
overhang (the G allele), and the second band corresponds to
molecules filled in with ddGTP at position 3 complementary to the
overhang (the A allele). The two bands were separated based on the
differences in molecular weight using gel electrophoresis. One
fluorescently labeled nucleotide ddGTP was used to determine that
an individual was heterozygous at a SNP site. This is the first use
of a single nucleotide to effectively detect the presence of two
different alleles.
[0630] For SNP TSC0034767, the template DNA in lanes 1 and 3 is
heterozygous for cytosine and guanine, as evidenced by the two
distinct bands. The lower band corresponded to ddGTP filled in at
position 1 complementary to the overhang. The second band of
slightly higher molecular weight corresponded to ddGTP filled in at
position 3, indicating that the first position in the overhang was
filled in with unlabeled dCTP, which allowed the polymerase to
continue to incorporate nucleotides until it incorporated ddGTP at
position 3 complementary to the overhang. The template DNA in lanes
2 and 4 was homozygous for guanine, as evidenced by a single band
of higher molecular weight than if ddGTP had been filled in at the
first position complementary to the overhang.
[0631] For SNP TSC1130902, the template DNA in lanes 1, 2, and 4 is
homozygous for adenine at the variable site, as evidenced by a
single higher molecular weight band migrating at about position 62
on the gel. The template DNA in lane 3 is heterozygous at the
variable site, as indicated by the presence of two distinct bands.
The lower band corresponds to molecules filled in with ddGTP at
position 1 complementary to the overhang (the guanine allele). The
higher molecular weight band corresponds to molecules filled in
with ddGTP at position 3 complementary to the overhang (the adenine
allele).
[0632] For SNP TSC0597888, the template DNA in lanes 1 and 4 was
homozygous for cytosine at the variable site; the template DNA in
lane 2 was heterozygous at the variable site, and the template DNA
in lane 3 was homozygous for guanine. The expected fill-in
reactions are diagrammed below:
[0633] Homozygous for Cytosine:
TABLE-US-00077 Allele 1 T C T G ATTC 3' G* A C TAAG 5' 4 3 2 1
Overhang position Allele 2 T C T G ATTC 3' G* A C TAAG 5' 4 3 2 1
Overhang position
[0634] Homozygous for Guanine:
TABLE-US-00078 Allele 1 T C T C ATTC 3' G* TAAG 5' 4 3 2 1 Overhang
position Allele 2 T C T C ATTC 3' G* TAAG 5' 4 3 2 1 Overhang
position
[0635] Heterozygous for Guanine/Cytosine:
TABLE-US-00079 Allele 1 T C T G ATTC 3' G* A C TAAG 5' 4 3 2 1
Overhang position Allele 2 T C T C ATTC 3' G* TAAG 5' 4 3 2 1
Overhang position
[0636] Template DNA homozygous for guanine at the variable site
displayed a single band, which corresponded to the DNA molecules
filled in with ddGTP at position 1 complementary to the overhang.
These DNA molecules were of lower molecular weight compared to the
DNA molecules filled in with ddGTP at position 3 of the overhang
(see lane 3 for SNP TSC0597888). The DNA molecules differed by two
bases in molecular weight.
[0637] Template DNA homozygous for cytosine at the variable site
displayed a single band, which corresponds to the DNA molecules
filled in with ddGTP at position 3 complementary to the overhang.
These DNA molecules migrated at a higher molecular weight than DNA
molecules filled in with ddGTP at position 1 (see lanes 1 and 4 for
SNP TSC0597888).
[0638] Template DNA heterozygous at the variable site displayed two
bands; one band corresponded to the DNA molecules filled in with
ddGTP at position 1 complementary to the overhang and was of lower
molecular weight, and the second band corresponded to DNA molecules
filled in with ddGTP at position 3 complementary to the overhang,
and was of higher molecular weight (see lane 3 for SNP
TSC0597888).
[0639] For SNP TSC0195492, the template DNA in lanes 1 and 3 was
heterozygous at the variable site, which was demonstrated by the
presence of two distinct bands. The template DNA in lane 2 was
homozygous for guanine at the variable site. The template DNA in
lane 4 was homozygous for cytosine. Only one band was seen in lane
4 for this SNP, and it had a higher molecular weight than the DNA
molecules filled in with ddGTP at position 1 complementary to the
overhang (compare lanes 2, 3 and 4).
[0640] The observed alleles for SNP TSC0607185 are reported as
cytosine or thymidine. For consistency, the SNP consortium denotes
the observed alleles as they appear in the sense strand
www.snp.cshl.org/shpsearch.shtml; website active as of Feb. 11,
2003). For this SNP, the second primer annealed upstream of the
locus of interest, which allowed the fill-in reaction to occur on
the antisense strand after digestion with BsmF I.
[0641] The template DNA in lanes 1 and 3 was heterozygous; the
template DNA in lane 2 was homozygous for thymidine, and the
template DNA in lane 4 was homozygous for cytosine. The antisense
strand was filled in with ddGTP, so the nucleotide on the sense
strand corresponded to cytosine.
[0642] Molecular weight markers can be used to identify the
positions of the expected bands. Alternatively, for each SNP
analyzed, a known heterozygous sample can be used, which will
identify precisely the position of the two expected bands.
[0643] As demonstrated in FIG. 11, one nucleotide labeled with one
fluorescent dye can be used to determine the identity of a variable
site including but not limited to SNPs and single nucleotide
mutations. Typically, to determine if an individual is homozygous
or heterozygous at a SNP site, multiple reactions are performed
using one nucleotide labeled with one dye and a second nucleotide
labeled with a second dye. However, this introduces problems in
comparing results because the two dyes have different quantum
coefficients. Even if different nucleotides are labeled with the
same dye, the quantum coefficients are different. The use of a
single nucleotide labeled with one dye eliminates any errors from
the quantum coefficients of different dyes.
[0644] In this example, fluorescently labeled ddGTP was used.
However, the method is applicable for a nucleotide tagged with any
signal generating moiety including but not limited to radioactive
molecule, fluorescent molecule, antibody, antibody fragment,
hapten, carbohydrate, biotin, derivative of biotin, phosphorescent
moiety, luminescent moiety, electrochemiluminescent moiety,
chromatic moiety, and moiety having a detectable electron spin
resonance, electrical capacitance, dielectric constant or
electrical conductivity. In addition, labeled ddATP, ddTTP, or
ddCTP can be used.
[0645] The above example used the third position complementary to
the overhang as an indicator of the second allele. However, the
second or fourth position of the overhang can be used as well (see
Section on Incorporation of Nucleotides). Furthermore, the overhang
was generated with the type IIS enzyme BsmF I; however any enzymes
that cuts DNA at a distance from its binding site can be used
including but not limited to the enzymes listed in Table I.
[0646] Also, in the above example, the nucleotide immediately
preceding the SNP site was not a guanine on the strand that was
filled in. This eliminated any effects of the alternative cutting
properties of the type IIS restriction enzyme to be removed. For
example, at SNP TSC0837969, the nucleotide from the SNP site on the
sense strand was an adenine. If BsmF I displayed alternate cutting
properties, the following overhangs would be generated for the
adenine allele and the guanine allele:
TABLE-US-00080 G allele - 11/15 Cut 5' TTA 3' AAT T C A C Overhang
position 0 1 2 3 G allele after fill-in 5' TTA A G* 3' AAT T C A C
Overhang position 0 1 2 3 A allele 11/15 Cut 5' TTA 3' AAT T T A C
Overhang position 0 1 2 3 A allele after fill-in 5' TTA A A T G* 3'
AAT T T A C Overhang position 0 1 2 3
[0647] For the guanine allele, the first position in the overhang
would be filled in with dATP, which would allow the polymerase to
incorporate ddGTP at position 2 complementary to the overhang.
There would be no detectable difference between molecules cut at
the 10/14 position or molecules cut at the 11/15 position.
[0648] For the adenine allele, the first position complementary to
the overhang would be filled in with dATP, the second position
would be filled in with dATP, the third position would be filled in
with dTTP, and the fourth position would be filled in with ddGTP.
There would be no difference in the molecular weights between
molecules cut at 10/14 or molecules cut at 11/15. The only
differences would correspond to whether the DNA molecules contained
an adenine at the variable site or a guanine at the variable
site.
[0649] As seen in FIG. 11, positioning the annealing region of the
first primer allows multiple SNPs to be analyzed in a single lane
of a gel. Also, when using the same nucleotide with the same dye, a
single fill-in reaction can be performed. In this example, 6 SNPs
were analyzed in one lane. However, any number of SNPs including
but not limited to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30,
30-40, 40-50, 51-60, 61-70, 71-80, 81-100, 101-120, 121-140,
141-160, 161-180, 181-200, and greater than 200 can be analyzed in
a single reaction.
[0650] Furthermore, one labeled nucleotide used to detect both
alleles can be mixed with a second labeled nucleotide used to
detect a different set of SNPs provided that neither of the
nucleotides that are labeled occur immediately before the variable
site (complementary to nucleotide at position 0 of the 11/15 cut).
For example, suppose SNP X can be guanine or thymidine at the
variable site and has the following 5' overhang generated after
digestion with BsmF I:
TABLE-US-00081 SNP X 10/14 5' TTGAC G allele 3'AACTG C A C T
Overhang position 1 2 3 4 SNP X 11/15 5' TTGA G allele 3'AACT G C A
C Overhang position 0 1 2 3 SNP X 10/14 5' TTGAC T allele 3'AACTG A
A C T Overhang position 1 2 3 4 SNP X 11/15 5' TTGA T allele 3'AACT
G A A C Overhang position 0 1 2 3
[0651] After the fill-in reaction with labeled ddGTP, unlabeled
dATP, dCTP, and dTTP, the following molecules would be
generated:
TABLE-US-00082 SNP X 10/14 5' TTGAC G* G allele 3'AACTG C A C T
Overhang position 1 2 3 4 SNP X 11/15 5' TTGA C G* G allele 3'AACT
G C A C Overhang position 0 1 2 3 SNP X 10/14 5' TTGAC T T G* T
allele 3'AACTG A A C T Overhang position 1 2 3 4 SNP X 11/15 5'
TTGA C T T G* T allele 3'AACT G A A C Overhang position 0 1 2 3
[0652] Now suppose SNP Y can be adenine or thymidine at the
variable site, and has the following 5' overhangs generated after
digestion with BsmF I.
TABLE-US-00083 SNP Y 10/14 5' GTTT A allele 3' CAAA T G T A
Overhang position 1 2 3 4 SNP Y 11/15 5' GTT A allele 3' CAA A T G
T Overhang position 0 1 2 3 SNP Y 10/14 5' GTTT T allele 3' CAAA A
G T A Overhang position 1 2 3 4 SNP Y 11/15 5' GTT T allele 3' CAA
A A G T Overhang position 0 1 2 3
[0653] After fill-in with labeled ddATP and unlabeled dCTP, dGTP,
and dTTP, the following molecules would be generated:
TABLE-US-00084 SNP Y 10/14 5' GTTT A* A allele 3' CAAA T G T A
Overhang position 1 2 3 4 SNP Y 11/15 5' GTTT T A* A allele 3' CAA
A T G T Overhang position 0 1 2 3 SNP Y 10/14 5' GTTT T C A* T
allele 3' CAAA A G T A Overhang position 1 2 3 4 SNP Y 11/15 5' GTT
T T C A* T allele 3' CAA A A G T Overhang position 0 1 2 3
[0654] In this example, labeled ddGTP and labeled ddATP are used to
determine the identity of both alleles of SNP X and SNP Y
respectively. The nucleotide immediately preceding (the
complementary nucleotide to position 0 of the overhang from the
11/15 cut SNP X is not guanine or adenine on the strand that is
filled-in. Likewise, the nucleotide immediately preceding SNPY is
not guanine or adenine on the strand that is filled-in. This allows
the fill-in reaction for both SNPs to occur in a single reaction
with labeled ddGTP, labeled ddATP, and unlabeled dCTP and dTTP.
This reduces the number of reactions that need to be performed and
increases the number of SNPs that can be analyzed in one
reaction.
[0655] The first primers for each SNP can be designed to anneal at
different distances from the locus of interest, which allows the
SNPs to migrate at different positions on the gel. For example, the
first primer used to amplify SNP X can anneal at 30 bases from the
locus of interest, and the first primer used to amplify SNP Y can
anneal at 35 bases from the locus of interest. Also, the
nucleotides can be labeled with fluorescent dyes that emit at
spectrums that do not overlap. After running the gel, the gel can
be scanned at one wavelength specific for one dye. Only those
molecules labeled with that dye will emit a signal. The gel then
can be scanned at the wavelength for the second dye. Only those
molecules labeled with that dye will emit a signal. This method
allows maximum compression for the number of SNPs that can be
analyzed in a single reaction.
[0656] In this example, the nucleotide preceding the variable site
on the strand that was filled-in was not adenine or guanine, and
the nucleotide following the variable site can not be adenine or
guanine on the sense strand. This method can work with any
combination of labeled nucleotides, and the skilled artisan would
understand which labeling reactions can be mixed and those that can
not. For instance, if one SNP is labeled with thymidine and a
second SNP is labeled with cytosine, the SNPs can be labeled in a
single reaction if the nucleotide immediately preceding each
variable site is not thymidine or cytosine on the sense strand and
the nucleotide immediately after the variable site is not thymidine
or cytosine on the sense strand.
[0657] This method allows the signals from one allele to be
compared to the signal from a second allele without the added
complexity of determining the degree of alternate cutting, or
having to correct for the quantum coefficients of the dyes. This
method is especially useful when trying to quantitate a ratio for
one allele to another. For example, this method is useful for
detecting chromosomal abnormalities. The ratio of alleles at a
heterozygous site is expected to be about 1:1 (one A allele and one
G allele). However, if an extra chromosome is present the ratio is
expected to be about 1:2 (one A allele and 2 G alleles or 2 A
alleles and 1 G allele). This method is especially useful when
trying to detect fetal DNA in the presence of maternal DNA.
[0658] In addition, this method is useful for detecting two genetic
signals in one sample. For example, this method can detect mutant
cells in the presence of wild type cells (see Example 5). If a
mutant cell contains a mutation in the DNA sequence of a particular
gene, this method can be used to detect both the mutant signal and
the wild type signal. This method can be used to detect the mutant
DNA sequence in the presence of the wild type DNA sequence. The
ratio of mutant DNA to wild type DNA can be quantitated because a
single nucleotide labeled with one signal generating moiety is
used.
Example 7
[0659] Non-invasive methods for the detection of various types of
cancer have the potential to reduce morbidity and mortality from
the disease. Several techniques for the early detection of
colorectal tumors have been developed including colonoscopy, barium
enemas, and sigmoidoscopy; however the techniques are limited in
use because they are invasive, which causes a low rate of patient
compliance. Non-invasive genetic tests may be useful in identifying
early stage colorectal tumors.
[0660] In 1991, researchers identified the Adenomatous Polyposis
Coli gene (APC), which plays a critical role in the formation of
colorectal tumors (Kinzler et al., Science 253:661-665, 1991). The
APC gene resides on chromosome 5q21-22 and a total of 15 exons code
for an RNA molecule of 8529 nucleotides, which produces a 300 Kd
APC protein. The protein is expressed in numerous cell types and is
essential for cell adhesion.
[0661] Mutations in the APC gene generally initiate colorectal
neoplasia (Tsao, J. et al., Am, J. Pathol. 145:531-534, 1994).
Approximately 95% of the mutations in the APC gene result in
nonsense/frameshift mutations. The most common mutations occur at
codons 1061 and 1309; mutations at these codons account for 1/3 of
all germline mutations. With regard to somatic mutations, 60% occur
within codons 1286-1513, which is about 10% of the coding sequence.
This region is termed the mutation Cluster Region (MCR), Numerous
types of mutations have been identified in the APC gene including
nucleotide substitutions (see Table VII), splicing errors (see
Table VIII), small deletions (see Table IX), small insertions (see
Table X), small insertions/deletions (see Table XI), gross
deletions (see Table XII), gross insertions (see Table XIII), and
complex rearrangements (see Table XIV).
[0662] Researchers have attempted to identify cells harboring
mutations in the APC gene in stool samples (Traverso, G. et al.,
New England Journal of Medicine, Vol 346:311-320, 2002). While APC
mutations are found in nearly all tumors, about 1 in 250 cells in
the stool sample has a mutation in the APC gene; most of the cells
are normal cells that have been shed into the feces. Furthermore,
human DNA represents about one-billionth of the total DNA found in
stool samples; the majority of DNA is bacterial. The technique
employed by Traverso et al. only detects mutations that result in a
truncated protein.
[0663] As discussed above, numerous mutations in the APC gene have
been implicated in the formation of colorectal tumors. Thus, a need
still exists for a highly sensitive, non-invasive technique for the
detection of colorectal tumors. Below, methods are described for
detection of two mutations in the APC gene. However, any number of
mutations can be analyzed using the methods described herein.
Preparation of Template DNA
[0664] The template DNA is purified from a sample containing colon
cells including but not limited to a stool sample. The template DNA
is purified using the procedures described by Ahlquist et al.
(Gastroenterology, 119:1219-1227, 2000). If stool samples are
frozen, the samples are thawed at room temperature, and homogenized
with an Exactor stool shaker (Exact Laboratories, Maynard, Mass.)
Following homogenization, a 4 gram stool equivalent of each sample
is centrifuged at 2536.times.g for 5 minutes. The samples are
centrifuged a second time at 16, 500.times.g for 10 minutes.
Supernatants are incubated with 20 .mu.l of RNase (0.5 mg per
milliliter) for 1 hour at 37.degree. C. DNA is precipitated with
1/10 volume of 3 mol of sodium acetate per liter and an equal
volume of isopropanol. The DNA is dissolved in 5 ml of TRIS-EDTA
(0.01 mol of Tris per liter (pH 7.4) and 0.001 mole of EDTA per
liter.
Design of Primers
[0665] To determine if a mutation resides at codon 1370, the
following primers are used:
TABLE-US-00085 First primer: (SEQ ED NO: 42) 5'
GTGCAAAGGCCTGAATTCCCAGGCACAAAGCTGTTGAA 3' Second primer: (SEQ ID
NO: 43) 5' TGAAGCGAACTAGGGACTCAGGTGGACTT
[0666] The first primer contains a biotin tag at the extreme 5'
end, and the nucleotide sequence for the restriction enzyme EcoRI.
The second primer contains the nucleotide sequence for the
restriction enzyme BsmF I.
[0667] To determine if a small deletion exists at codon 1302, the
following primers are used:
TABLE-US-00086 First primer: (SEQ ID NO: 44) 5'
GATTCCGTAAACGAATTCAGTTCATTATCATCTTTGTC 3' Second primer: (SEQ ID
NO: 45) 5' CCATTGTTAAGCGGGACTTCTGCTATTTG 3'
[0668] The first primer has a biotin tag at the 5' end and contains
a restriction enzyme recognition site for EcoRI. The second primer
contains a restriction enzyme recognition site for BsmF I,
PCR Reaction
[0669] The loci of interest are amplified from the template genomic
DNA using the polymerase chain reaction (PCR, U.S. Pat. Nos.
4,683,195 and 4,683,202, incorporated herein by reference). The
loci of interest are amplified in separate reaction tubes; they can
also be amplified together in a single PCR reaction. For increased
specificity, a "hot-start" PCR reaction is used, e.g. by using the
HotStarTaq Master Mix Kit supplied by QIAGEN (catalog number
203443). The amount of template DNA and primer per reaction are
optimized for each locus of interest but in this example, 40 ng of
template human genomic DNA and 5 .mu.M of each primer are used.
Forty cycles of PCR are performed. The following PCR conditions are
used: [0670] (1) 95.degree. C. for 15 minutes and 15 seconds;
[0671] (2) 37.degree. C. for 30 seconds; [0672] (3) 95.degree. C.
for 30 seconds; [0673] (4) 57.degree. C. for 30 seconds; [0674] (5)
95.degree. C. for 30 seconds; [0675] (6) 64.degree. C. for 30
seconds; [0676] (7) 95.degree. C. for 30 seconds; [0677] (8) Repeat
steps 6 and 7 thirty nine (39) times; [0678] (9) 72.degree. C. for
5 minutes.
[0679] In the first cycle of PCR, the annealing temperature is
about the melting temperature of the 3' annealing region of the
second primers, which is 37.degree. C. The annealing temperature in
the second cycle of PCR is about the melting temperature of the 3'
region, which anneals to the template DNA, of the first primer,
which is 57.degree. C. The annealing temperature in the third cycle
of PCR is about the melting temperature of the entire sequence of
the second primer, which is 64.degree. C. The annealing temperature
for the remaining cycles is 64.degree. C. Escalating the annealing
temperature from TM1 to TM2 to TM3 in the first three cycles of PCR
greatly improves specificity. These annealing temperatures are
representative, and the skilled artisan understands that the
annealing temperatures for each cycle are dependent on the specific
primers used.
[0680] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
Purification of Fragment of Interest
[0681] The PCR products are separated from the genomic template
DNA. Each PCR product is divided into four separate reaction wells
of a Streptawell, transparent, High-Bind plate from Roche
Diagnostics GmbH (catalog number 1 645 692, as listed in Roche
Molecular Biochemicals, 2001 Biochemicals Catalog). The first
primers contain a 5' biotin tag so the PCR products bound to the
Streptavidin coated wells while the genomic template DNA does not.
The streptavidin binding reaction is performed using a Thermomixer
(Eppendorf) at 1000 rpm for 20 min. at 37.degree. C. Each well is
aspirated to remove unbound material, and washed three times with
1.times.PBS, with gentle mixing (Kandpal et al., Nucl. Acids Res.
18:1789-1795 (1990); Kaneoka et al., Biotechniques 10:30-34 (1991);
Green et al., Nucl. Acids Res. 18:6163-6164 (1990)).
[0682] Alternatively, the PCR products are placed into a single
well of a streptavidin plate to perform the nucleotide
incorporation reaction in a single well.
Restriction Enzyme Digestion of Isolated Fragments
[0683] The purified PCR products are digested with the restriction
enzyme BsmF I (New England Biolabs catalog number R0572S), which
binds to the recognition site incorporated into the PCR products
from the second primer. The digests are performed in the
Streptawells following the instructions supplied with the
restriction enzyme. After digestion with the appropriate
restriction enzyme, the wells are washed three times with PBS to
remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[0684] The restriction enzyme digest described above yields a DNA
fragment with a 5' overhang, which contains the locus of interest
and a 3' recessed end. The 5' overhang functions as a template
allowing incorporation of a nucleotide or nucleotides in the
presence of a DNA polymerase.
[0685] For each locus of interest, four separate fill in reactions
are performed; each of the four reactions contains a different
fluorescently labeled ddNTP (ddATP, ddTTP, ddGTP, or ddCTP). The
following components are added to each fill in reaction: 1 .mu.l of
a fluorescently labeled ddNTP, 0.5 .mu.l of unlabeled ddNTPs (40
.mu.M), which contains all nucleotides except the nucleotide that
is fluorescently labeled, 2 .mu.l of 10.times. sequenase buffer,
025 .mu.l of Sequenase, and water as needed for a 20 .mu.l
reaction. The fill are performed in reactions at 40.degree. C. for
10 min. Non-fluorescently labeled ddNTP are purchased from
Fermentas Inc. (Hanover, Md.). All other labeling reagents are
obtained from Amersham (Thermo Sequenase Dye Terminator Cycle
Sequencing Core Kit, US 79565). In the presence of fluorescently
labeled ddNTPs, the 3' recessed end is extended by one base, which
corresponds to the locus of interest.
[0686] A mixture of labeled ddNTPs and unlabeled dNTPs also can be
used for the fill-in reaction. The "fill in" conditions are as
described above except that a mixture containing 40 .mu.M unlabeled
dNTPs, 1 .mu.l fluorescently labeled ddATP, 1 .mu.l fluorescently
labeled ddTTP, 1 .mu.l fluorescently labeled ddCTP, and 1 .mu.l
ddGTP are used. The fluorescent ddNTPs are obtained from Amersham
(Thermo Sequenase Dye Terminator Cycle Sequencing Core Kit, US
79565; Amersham does not publish the concentrations of the
fluorescent nucleotides). The locus of interest is digested with
the restriction enzyme BsmF I, which generates a 5' overhang of
four bases. If the first nucleotide incorporated is a labeled
ddNTP, the 3' recessed end is filled in by one base, allowing
detection of the locus of interest. However, if the first
nucleotide incorporated is a dNTP, the polymerase continues to
incorporate nucleotides until a ddNTP is filled in. For example,
the first two nucleotides may be filled in with dNTPs, and the
third nucleotide with a ddNTP, allowing detection of the third
nucleotide in the overhang. Thus, the sequence of the entire 5'
overhang is determined, which increases the information obtained
from each SNP or locus of interest. This type of fill in reaction
is especially useful when detecting the presence of insertions,
deletions, insertions and deletions, rearrangements, and
translocations.
[0687] Alternatively, one nucleotide labeled with a single dye is
used to determine the sequence of the locus of interest. See
Example 6. This method eliminates any potential errors when using
different dyes, which have different quantum coefficients.
[0688] After labeling, each Streptawell is rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments are released
from the Streptawells by digesting with the restriction enzyme
EcoRI, according to the manufacturer's instructions that are
supplied with the enzyme. The digestion is performed for 1 hour at
37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[0689] After release from the streptavidin matrix, the sample is
loaded into a lane of a 36 cm 5% acrylamide (urea) gel
(BioWhittaker Molecular Applications, Long Ranger Run Gel Packs,
catalog number 50691). The sample is electrophoresed into the gel
at 3000 volts for 3 min. The gel is run for 3 hours using a
sequencing apparatus (Hoefer SQ3 Sequencer). The incorporated
labeled nucleotide is detected by fluorescence.
[0690] To determine if any cells contain mutations at codon 1370 of
the APC gene when separate fill-in reactions are performed, the
lanes of the gel that correspond to the fill-in reaction for ddATP
and ddTTP are analyzed. If only normal cells are present, the lane
corresponding to the fill in reaction with ddATP is a bright
signal. No signal is detected for the "fill-in" reaction with
ddTTP. However, if the patient sample contains cells with mutations
at codon 1370 of the APC gene, the lane corresponding to the fill
in reaction with ddATP is a bright signal, and a signal is detected
from the lane corresponding to the fill in reaction with ddTTP. The
intensity of the signal from the lane corresponding to the fill in
reaction with ddTTP is indicative of the number of mutant cells in
the sample.
[0691] Alternatively, one labeled nucleotide is used to determine
the sequence of the alleles at codon 1370 of the APC gene. At codon
1370, the normal sequence is AAA, which codes for the amino acid
lysine. However, a nucleotide substitution has been identified at
codon 1370, which is associated with colorectal tumors.
Specifically, a change from A to T (AAA-TAA) typically is found at
codon 1370, which results in a stop codon. A single fill-in
reaction is performed using labeled ddATP, and unlabeled dTTP,
dCTP, and dGTP. A single nucleotide labeled with one fluorescent
dye is used to determine the presence of both the normal and mutant
DNA sequence that codes for codon 1370. The relevant DNA sequence
is depicted below with the sequence corresponding to codon 1370 in
bold:
TABLE-US-00087 5' CCCAAAAGTCCACCTGA (SEQ ID NO: 46) 3'
GGGTTTTCAGGTGGACT (SEQ ID NO: 47)
[0692] After digest with BsmF I, the following overhang is
produced:
TABLE-US-00088 5' CCC 3' GGG T T T T Overhang position 1 2 3 4
[0693] If the patient sample has no cells harboring a mutation at
codon 1370, one signal is seen corresponding to incorporation of
labeled ddATP.
TABLE-US-00089 5' CCC A* 3' GGG T T T T Overhang position 1 2 3
4
[0694] However, if the patient sample has cells with mutations at
codon 1370 of the APC gene, one signal is seen, which corresponds
to the normal sequence at codon 1370, and a second signal is seen,
which corresponds to the mutant sequence at codon 1370. The signals
clearly are identified as they differ in molecular weight.
TABLE-US-00090 Overhang of normal DNA sequence: CCC GGG T T T T
Overhang position 1 2 3 4 Normal DNA sequence after fill-in: CCC A*
GGG T T T T Overhang position 1 2 3 4 Overhang of mutant DNA
sequence: CCC GGG A T T T Overhang position 1 2 3 4 Mutant DNA
sequence after fill-in: CCC T A* GGG A T T T Overhang position 1 2
3 4
[0695] Two signals are seen when the mutant allele is present. The
mutant DNA molecules are filled in one base after the wild type DNA
molecules. The two signals are separated using any method that
discriminates based on molecular weight. One labeled nucleotide
(ddATP) is used to detect the presence of both the wild type DNA
sequence and the mutant DNA sequence. This method of labeling
reduces the number of reactions that need to be performed and
allows accurate quantitation for the number of mutant cells in the
patient sample. The number of mutant cells in the sample is used to
determine patient prognosis, the degree and the severity of the
disease. This method of labeling eliminates the complications
associated with using different dyes, which have distinct quantum
coefficients. This method of labeling also eliminates errors
associated with pipetting reactions.
[0696] To determine if any cells contain mutations at codon 1302 of
the APC gene when separate fill-in reactions are performed, the
lanes of the gel that correspond to the fill-in reaction for ddTTP
and ddCTP are analyzed. The normal DNA sequence is depicted below
with sequence coding for codon 1302 in bold type-face.
TABLE-US-00091 Normal Sequence: 5' ACCCTGCAAATAGCAGAA (SEQ ID NO:
48) 3' TGGGACGTTTATCGTCTT (SEQ ID NO: 49)
[0697] After digest, the following 5' overhang is produced:
TABLE-US-00092 5' ACCC 3' TGGG A C G T Overhang position 1 2 3
4
[0698] After the fill-in reaction, labeled ddTTP is
incorporated.
TABLE-US-00093 5' ACCC T* 3' TGGG A C G T Overhang position 1 2 3
4
[0699] A deletion of a single base of the APC sequence, which
typically codes for codon 1302, has been associated with colorectal
tumors. The mutant DNA sequence is depicted below with the relevant
sentience in bold:
TABLE-US-00094 Mutant Sequence: 5' ACCCGCAAATAGCAGAA (SEQ ID NO:
50) 3' TGGGCGTTTATCGTCTT (SEQ ID NO: 51)
[0700] After Digest:
TABLE-US-00095 5' ACC 3' TGG G C G T Overhang position 1 2 3 4
[0701] After Fill-in:
TABLE-US-00096 5' ACC C* 3' TGG G C G T Overhang position 1 2 3
4
[0702] If there are no mutations in the APC gene, signal is not
detected for the fill in reaction with ddCTP*, but a bright signal
is detected for the fill-in reaction with ddTTP*. However, if there
are cells in the patient sample that have mutations in the APC
gene, signals are seen for the fill-in reactions with ddCTP*and
ddTTP*.
[0703] Alternatively, a single fill-in reaction is performed using
a mixture containing unlabeled dNTPs, fluorescently labeled ddATP,
fluorescently labeled ddTTP, fluorescently labeled ddCTP, and
fluorescently labeled ddGTP. If there is no deletion, labeled ddTTP
is incorporated.
TABLE-US-00097 5' ACCC T* 3' TGGG A C G T Overhang position 1 2 3
4
[0704] However, if the T has been deleted, labeled ddCTP* is
incorporated.
TABLE-US-00098 5' ACCC* 3' TGGG C G T Overhang position 1 2 3 4
[0705] The two signals are separated by molecular weight because of
the deletion of the thymidine nucleotide. If mutant cells are
present, two signals are generated in the same lane but are
separated by a single base pair (this principle is demonstrated in
FIG. 9D). The deletion causes a change in the molecular weight of
the DNA fragments, which allows a single fill in reaction to be
used to detect the presence of both normal and mutant cells.
[0706] In the above example, methods for the detection of a
nucleotide substitution and a small deletion are described.
However, the methods can be used for the detection of any type of
mutation including but not limited to nucleotide substitutions (see
Table VII), splicing errors (see Table VIII), small deletions (see
Table IX), small insertions (see Table X), small
insertions/deletions (see Table XI), gross deletions (see Table
XII), gross insertions (see Table XIII), and complex rearrangements
(see Table XIV).
[0707] In addition, the above-described methods are used for the
detection of any type of disease including but not limited to those
listed in Table IV. Furthermore, any type of mutant gene is
detected using the inventions described herein including but not
limited to the genes associated with the diseases listed in Table
IV, BRCA1, BRCA2, MSH6, MSH2, MLH1, RET, PTEN, ATM, H-RAS, p53,
ELAC2, CDH1, APC, AR, PMS2, MLH3, CYP1A1, GSTP1, GSTM1, AXIN2,
CYP19, MET, NAT1, CDKN2A, NQ01, trc8, RAD51, PMS1, TGFBR2, VHL,
MC4R, POMC, NROB2, UCP2, PCSK1, PPARG, ADRB2, UCP3, glur1, cart,
SORBS1, LEP, LEPR, SIM1, TNF, IL-6, IL-1, IL-2, IL-3, IL1A, TAP2,
THPO, THRB, NBS1, RBM15, LIF, MPL, RUNX1, Her-2, glucocorticoid
receptor, estrogen receptor, thyroid receptor, p21, p27, K-RAS,
N-RAS, retinoblastoma protein, Wiskott-Aldrich (WAS) gene, Factor V
Leiden, Factor II (prothrombin), methylene tetrahydrofolate
reductase, cystic fibrosis, LDL receptor, HDL receptor, superoxide
dismutase gene, SHOX gene, genes involved in nitric oxide
regulation, genes involved in cell cycle regulation, tumor
suppressor genes, oncogenes, genes associated with
neurodegeneration, genes associated with obesity. Abbreviations
correspond to the proteins as listed on the Human Gene Mutation
Database, which is incorporated herein by reference
www.archive.uwcm.ac.uk./uwcm; website address active as of Feb. 12,
2003).
[0708] The above example demonstrates the detection of mutant cells
and mutant alleles from a fecal sample. However, the methods
described herein are used for detection of mutant cells from any
biological sample including but not limited to blood sample, serum
sample, plasma sample, urine sample, spinal fluid, lymphatic fluid,
semen, vaginal secretion, ascitic fluid, saliva, mucosa secretion,
peritoneal fluid, fecal sample, body exudates, breast fluid, lung
aspirates, cells, tissues, individual cells or extracts of the such
sources that contain the nucleic acid of the same, and subcellular
structures such as mitochondria or chloroplasts. In addition, the
methods described herein are used for the detection of mutant cells
and mutated DNA from any number of nucleic acid containing sources
including but not limited to forensic, food, archeological,
agricultural or inorganic samples.
[0709] The above example is directed to detection of mutations in
the APC gene. However, the inventions described herein are used for
the detection of mutations in any gene that is associated with or
predisposes to disease (see Table XV).
[0710] For example, hypermethylation of the glutathione
S-transferase P1 (GSTP1) promoter is the most common DNA alteration
in prostrate cancer. The methylation state of the promoter is
determined using sodium bisulfite and the methods described
herein.
[0711] Treatment with sodium bisulfite converts unmethylated
cytosine residues into uracil, and leaving the methylated cytosines
unchanged. Using the methods described herein, a first and second
primer are designed to amplify the regions of the GSTP1 promoter
that are often methylated. Below, a region of the GSTP1 promoter is
shown prior to sodium bisulfite treatment:
[0712] Before Sodium Bisulfite Treatment:
TABLE-US-00099 5' ACCGCTACA 3' TGGCGATCA
[0713] Below, a region of the GSTP1 promoter is shown after sodium
bisulfite treatment, PCR amplification, and digestion with the type
IIS restriction enzyme BsmF I;
TABLE-US-00100 Unmethylated 5' ACC 3' TGG U G A T Overhang position
1 2 3 4 Methylated 5' ACC 3' TGG C G A T Overhang position 1 2 3
4
[0714] Labeled ddATP, unlabeled dCTP, dGTP, and dTTP are used to
fill-in the 5' overhangs. The following molecules are
generated:
TABLE-US-00101 Unmethylated 5' ACC A* 3' TGG U G A T Overhang
position 1 2 3 4 Methylated 5' ACC G C T A* 3' TGG C G A T Overhang
position 1 2 3 4
[0715] Two signals are seen; one corresponds to DNA molecules
filled in with ddATP at position one complementary to the overhang
(unmethylated), and the other corresponds to the DNA molecules
filled in with ddATP at position 4 complementary to the overhang
(methylated). The two signals are separated based on molecular
weight. Alternatively, the fill-in reactions are performed in
separate reactions using labeled ddGTP in one reaction and labeled
ddATP in another reaction.
[0716] The methods described herein are used to screen for prostate
cancer and also to monitor the progression and severity of the
disease. The use of a single nucleotide to detect both the
methylated and unmethylated sequences allows accurate quantitation
and provides a high level of sensitivity for the methylated
sequences, which is a useful tool for earlier detection of the
disease.
[0717] The information contained in Tables VII-XIV was obtained
from the Human Gene Mutation Database. With the information
provided herein, the skilled artisan will understand how to apply
these methods for determining the sequence of the alleles for any
gene. A large number of genes and there associated mutations can be
found at the following website: www.archive.uwcm.ac.uk./uwcm.
TABLE-US-00102 TABLE VII NUCLEOTIDE SUBSTITUTIONS Co- Amino don
Nucleotide acid Phenotype 99 CGG-TGG Arg-Trp Adenomatous polyposis
coli 121 AGA-TGA Arg-Term Adenomatous polyposis coli 157 TGG-TAG
Trp-Term Adenomatous polyposis coli 159 TAC-TAG Tyr-Term
Adenomatous polyposis coli 163 CAG-TAG Gln-Term Adenomatous
polyposis coli 168 AGA-TGA Arg-Term Adenomatous polyposis coli 171
AGT-ATT Ser-Ile Adenomatous polyposis coli 181 CAA-TAA Gln-Term
Adenomatous polyposis coli 190 GAA-TAA Glu-Term Adenomatous
polyposis coli 202 GAA-TAA Glu-Term Adenomatous polyposis coli 208
CAG-CGG Gln-Arg Adenomatous polyposis coli 208 CAG-TAG Gln-Term
Adenomatous polyposis coli 213 CGA-TGA Arg-Term Adenomatous
polyposis coli 215 CAG-TAG Gln-Term Adenomatous polyposis coli 216
CGA-TGA Arg-Term Adenomatous polyposis coli 232 CGA-TGA Arg-Term
Adenomatous polyposis coli 233 CAG-TAG Gln-Term Adenomatous
polyposis coli 247 CAG-TAG Gln-Term Adenomatous polyposis coli 267
GGA-TGA Gly-Term Adenomatous polyposis coli 278 CAG-TAG Gln-Term
Adenomatous polyposis coli 280 TCA-TGA Ser-Term Adenomatous
polyposis coli 280 TCA-TAA Ser-Term Adenomatous polyposis coli 283
CGA-TGA Arg-Term Adenomatous polyposis coli 302 CGA-TGA Arg-Term
Adenomatous polyposis coli 332 CGA-TGA Arg-Term Adenomatous
polyposis coli 358 CAG-TAG Gln-Term Adenomatous polyposis coli 405
CGA-TGA Arg-Term Adenomatous polyposis coli 414 CGC-TGC Arg-Cys
Adenomatous polyposis coli 422 GAG-TAG Glu-Term Adenomatous
polyposis coli 423 TGG-TAG Trp-Term Adenomatous polyposis coli 424
CAG-TAG Gln-Term Adenomatous polyposis coli 433 CAG-TAG Gln-Term
Adenomatous polyposis coli 443 GAA-TAA Glu-Term Adenomatous
polyposis coli 457 TCA-TAA Ser-Term Adenomatous polyposis coli 473
CAG-TAG Gln-Term Adenomatous polyposis coli 486 TAC-TAG Tyr-Term
Adenomatous polyposis coli 499 CGA-TGA Arg-Term Adenomatous
polyposis coli 500 TAT-TAG Tyr-Term Adenomatous polyposis coli 541
CAG-TAG Gln-Term Adenomatous polyposis coli 553 TGG-TAG Trp-Term
Adenomatous polyposis coli 554 CGA-TGA Arg-Term Adenomatous
polyposis coli 564 CGA-TGA Arg-Term Adenomatous polyposis coli 577
TTA-TAA Leu-Term Adenomatous polyposis coli 586 AAA-TAA Lys-Term
Adenomatous polyposis coli 592 TTA-TGA Leu-Term Adenomatous
polyposis coli 593 TGG-TAG Trp-Term Adenomatous polyposis coli 593
TGG-TGA Trp-Term Adenomatous polyposis coli 622 TAC-TAA Tyr-Term
Adenomatous polyposis coli 625 CAG-TAG Gln-Term Adenomatous
polyposis coli 629 TTA-TAA Leu-Term Adenomatous polyposis coli 650
GAG-TAG Glu-Term Adenomatous polyposis coli 684 TTG-TAG Leu-Term
Adenomatous polyposis coli 685 TGG-TGA Trp-Term Adenomatous
polyposis coli 695 CAG-TAG Gln-Term Adenomatous polyposis coli 699
TGG-TGA Trp-Term Adenomatous polyposis coli 699 TGG-TAG Trp-Term
Adenomatous polyposis coli 713 TCA-TGA Ser-Term Adenomatous
polyposis coli 722 AGT-GGT Ser-Gly Adenomatous polyposis coli 747
TCA-TGA Ser-Term Adenomatous polyposis coli 764 TTA-TAA Leu-Term
Adenomatous polyposis coli 784 TCT-ACT Ser-Thr Adenomatous
polyposis coli 805 CGA-TGA Arg-Term Adenomatous polyposis coli 811
TCA-TGA Ser-Term Adenomatous polyposis coli 848 AAA-TAA Lys-Term
Adenomatous polyposis coli 876 CGA-TGA Arg-Term Adenomatous
polyposis coli 879 CAG-TAG Gln-Term Adenomatous polyposis coli 893
GAA-TAA Glu-Term Adenomatous polyposis coli 932 TCA-TAA Ser-Term
Adenomatous polyposis coli 932 TCA-TGA Ser-Term Adenomatous
polyposis coli 935 TAC-TAG Tyr-Term Adenomatous polyposis coli 935
TAC-TAA Tyr-Term Adenomatous polyposis coli 995 TGC-TGA Cys-Term
Adenomatous polyposis coli 997 TAT-TAG Tyr-Term Adenomatous
polyposis coli 999 CAA-TAA Gln-Term Adenomatous polyposis coli 1000
TAC-TAA Tyr-Term Adenomatous polyposis coli 1020 GAA-TAA Glu-Term
Adenomatous polyposis coli 1032 TCA-TAA Ser-Term Adenomatous
polyposis coli 1041 CAA-TAA Gln-Term Adenomatous polyposis coli
1044 TCA-TAA Ser-Term Adenomatous polyposis coli 1045 CAG-TAG
Gln-Term Adenomatous polyposis coli 1049 TGG-TGA Trp-Term
Adenomatous polyposis coli 1067 CAA-TAA Gln-Term Adenomatous
polyposis coli 1071 CAA-TAA Gln-Term Adenomatous polyposis coli
1075 TAT-TAA Tyr-Term Adenomatous polyposis coli 1075 TAT-TAG
Tyr-Term Adenomatous polyposis coli 1102 TAC-TAG Tyr-Term
Adenomatous polyposis coli 1110 TCA-TGA Ser-Term Adenomatous
polyposis coli 1114 CGA-TGA Arg-Term Adenomatous polyposis coli
1123 CAA-TAA Gln-Term Adenomatous polyposis coli 1135 TAT-TAG
Tyr-Term Adenomatous polyposis coli 1152 CAG-TAG Gln-Term
Adenomatous polyposis coli 1155 GAA-TAA Glu-Term Adenomatous
polyposis coli 1168 GAA-TAA Glu-Term Adenomatous polyposis coli
1175 CAG-TAG Gln-Term Adenomatous polyposis coli 1176 CCT-CTT
Pro-Leu Adenomatous polyposis coli 1184 GCC-CCC Ala-Pro Adenomatous
polyposis coli 1193 CAG-TAG Gln-Term Adenomatous polyposis coli
1194 TCA-TGA Ser-Term Adenomatous polyposis coli 1198 TCA-TGA
Ser-Term Adenomatous polyposis coli 1201 TCA-TGA Ser-Term
Adenomatous polyposis coli 1228 CAG-TAG Gln-Term Adenomatous
polyposis coli 1230 CAG-TAG Gln-Term Adenomatous polyposis coli
1244 CAA-TAA Gln-Term Adenomatous polyposis coli 1249 TGC-TGA
Cys-Term Adenomatous polyposis coli 1256 CAA-TAA Gln-Term
Adenomatous polyposis coli 1262 TAT-TAA Tyr-Term Adenomatous
polyposis coli 1270 TGT-TGA Cys-Term Adenomatous polyposis coli
1276 TCA-TGA Ser-Term Adenomatous polyposis coli 1278 TCA-TAA
Ser-Term Adenomatous polyposis coli 1286 GAA-TAA Glu-Term
Adenomatous polyposis coli 1289 TGT-TGA Cys-Term Adenomatous
polyposis coli 1294 CAG-TAG Gln-Term Adenomatous polyposis coli
1307 ATA-AAA Ile-Lys Colorectal cancer, predisposition to,
association 1309 GAA-TAA Glu-Term Adenomatous polyposis coli 1317
GAA-CAA Glu-Gln Colorectal cancer, predisposition to 1328 CAG-TAG
Gln-Term Adenomatous polyposis coli 1338 CAG-TAG Gln-Term
Adenomatous polyposis coli 1342 TTA-TAA Leu-Term Adenomatous
polyposis coli 1342 TTA-TGA Leu-Term Adenomatous polyposis coli
1348 AGG-TGG Arg-Trp Adenomatous polyposis coli 1357 GGA-TGA
Gly-Term Adenomatous polyposis coli
1367 CAG-TAG Gln-Term Adenomatous polyposis coli 1370 AAA-TAA
Lys-Term Adenomatous polyposis coli 1392 TCA-TAA Ser-Term
Adenomatous polyposis coli 1392 TCA-TGA Ser-Term Adenomatous
polyposis coli 1397 GAG-TAG Glu-Term Adenomatous polyposis coli
1449 AAG-TAG Lys-Term Adenomatous polyposis coli 1450 CGA-TGA
Arg-Term Adenomatous polyposis coli 1451 GAA-TAA Glu-Term
Adenomatous polyposis coli 1503 TCA-TAA Ser-Term Adenomatous
polyposis coli 1517 CAG-TAG Gln-Term Adenomatous polyposis coli
1529 CAG-TAG Gln-Term Adenomatous polyposis coli 1539 TCA-TAA
Ser-Term Adenomatous polyposis coli 1541 CAG-TAG Gln-Term
Adenomatous polyposis coli 1564 TTA-TAA Leu-Term Adenomatous
polyposis coli 1567 TCA-TGA Ser-Term Adenomatous polyposis coli
1640 CGG-TGG Arg-Trp Adenomatous polyposis coli 1693 GAA-TAA
Glu-Term Adenomatous polyposis coli 1822 GAC-GTC Asp-Val
Adenomatous polyposis coli, association with ? 2038 CTG-GTG Leu-Val
Adenomatous polyposis coli 2040 CAG-TAG Gln-Term Adenomatous
polyposis coli 2566 AGA-AAA Arg-Lys Adenomatous polyposis coli 2621
TCT-TGT Ser-Cys Adenomatous polyposis coli 2839 CTT-TTT Leu-Phe
Adenomatous polyposis coli
TABLE-US-00103 TABLE VIII NUCLEOTIDE SUBSTITUTIONS Relative
Donor/Acceptor location Substitution Phenotype ds -1 G-C
Adenomatous polyposis coli as -1 G-A Adenomatous polyposis coli as
-1 G-C Adenomatous polyposis coli ds +2 T-A Adenomatous polyposis
coli as -1 G-C Adenomatous polyposis coli as -1 G-T Adenomatous
polyposis coli as -1 G-A Adenomatous polyposis coli as -2 A-C
Adenomatous polyposis coli as -5 A-G Adenomatous polyposis coli ds
+3 A-C Adenomatous polyposis coli as -1 G-A Adenomatous polyposis
coli ds +1 G-A Adenomatous polyposis coli as -1 G-T Adenomatous
polyposis coli ds +1 G-A Adenomatous polyposis coli as -1 G-A
Adenomatous polyposis coli ds +1 G-A Adenomatous polyposis coli ds
+3 A-G Adenomatous polyposis coli ds +5 G-T Adenomatous polyposis
coli as -1 G-A Adenomatous polyposis coli as -6 A-G Adenomatous
polyposis coli as -5 A-G Adenomatous polyposis coli as -2 A-G
Adenomatous polyposis coli ds +2 T-C Adenomatous polyposis coli as
-2 A-G Adenomatous polyposis coli ds +1 G-A Adenomatous polyposis
coli ds +1 G-T Adenomatous polyposis coli ds +2 T-G Adenomatous
polyposis coli
TABLE-US-00104 TABLE IX APC SMALL DELETIONS Location/ codon
Deletion Phenotype 77 TTAgataGCAGTAATTT Adenomatous SEQ ID NO: 52
polyposis coli 97 GGAAGccgggaagGATCTGTATC Adenomatous SEQ ID NO: 53
polyposis coli 138 GAGAaAGAGAG_E3I3_GTAA Adenomatous SEQ ID NO: 54
polyposis coli 139 AAAGAgag_E3I3_Gtaacttttct Thyroid cancer SEQ ID
NO: 55 139 AAAGagag_E3I3_GTAACTTTTC Adenomatous SEQ ID NO: 56
polyposis coli 142 TTTTAAAAAAaAAAAATAG_I3E4_GTCA Adenomatous SEQ ID
NO: 57 polyposis coli 144 AAAATAG_I3E4_GTCatTGCTTCTTGC Adenomatous
SEQ ID NO: 58 polyposis coli 149 GACAaaGAAGAAAAGG Adenomatous SEQ
ID NO: 59 polyposis coli 149 GACAAagaaGAAAAGGAAA Adenomatous SEQ ID
NO: 60 polyposis coli 155 AGGAA{circumflex over (
)}AAAGActggtATTACGCTCA Adenomatous SEQ ID NO: 61 polyposis coli 169
AAAAGA{circumflex over ( )}ATAGatagTCTTCCTTTA Adenomatous SEQ ID
NO: 62 polyposis coli 172 AGATAGT{circumflex over (
)}CTTcCTTTAACTGA Adenomatous SEQ ID NO: 63 polyposis coli 179
TCCTTacaaACAGATATGA Adenomatous SEQ ID NO: 64 polyposis coli 185
ACCaGAAGGCAATT Adenomatous SEQ ID NO: 65 polyposis coli 196
ATCAGagTTGCGATGGA Adenomatous SEQ ID NO: 66 polyposis coli 213
CGAGCaCAG_E5I5_GTAAGTT Adenomatous SEQ ID NO: 67 polyposis coli 298
CACtcTGCACCTCGA Adenomatous SEQ ID NO: 68 polyposis coli 329
GATaTGTCGCGAAC Adenomatous SEQ ID NO: 69 polyposis coli 365
AAAGActCTGTATTGTT Adenomatous SEQ ID NO: 70 polyposis coli 397
GACaaGAGAGGCAGG Adenomatous SEQ ID NO: 71 polyposis coli 427
CATGAacCAGGCATGGA Adenomatous SEQ ID NO: 72 polyposis coli 428
GAACCaGGCATGGACC Adenomatous SEQ ID NO: 73 polyposis coli 436
AATCCaa_E9I9_gTATGTTCTCT Adenomatous SEQ ID NO: 74 polyposis coli
440 GCTCCtGTTGAACATC Adenomatous SEQ ID NO: 75 polyposis coli 455
AAACTtTCATTTGATG Adenomatous SEQ ID NO: 76 polyposis coli 455
AAACtttcaTTTGATGAAG Adenomatous SEQ ID NO: 77 polyposis coli 472
CTAcAGGCCATTGC Adenomatous SEQ ID NO: 78 polyposis coli 472
TAAATTAG_I10E11_GGgGACTACAGGC Adenomatous SEQ ID NO: 79 polyposis
coli 478 TTATtGCAAGTGGAC Adenomatous SEQ ID NO: 80 polyposis coli
486 TACGgGCTTACTAAT Adenomatous SEQ ID NO: 81 polyposis coli 494
AGTATtACACTAAGAC Adenomatous SEQ ID NO: 82 polyposis coli 495
ATTACacTAAGACGATA Adenomatous SEQ ID NO: 83 polyposis coli 497
CTAaGACGATATGC Adenomatous SEQ ID NO: 84 polyposis coli 520
TGCTCtaTGAAAGGCTG Adenomatous SEQ ID NO: 85 polyposis coli 526
ATGAGagcacttgtgGCCCAACTAA Adenomatous SEQ ID NO: 86 polyposis coli
539 GACTTaCAGCAG_E12I12_GTAC Adenomatous SEQ ID NO: 87 polyposis
coli 560 AAAAAgaCGTTGCGAGA Adenomatous SEQ ID NO: 88 polyposis coli
566 GTTGgaagtGTGAAAGCAT Adenomatous SEQ ID NO: 89 polyposis coli
570 AAAGCaTTGATGGAAT Adenomatous SEQ ID NO: 90 polyposis coli 577
TTAGaagtTAAAAAG_E13I13_GTA Adenomatous SEQ ID NO: 91 polyposis coli
584 ACCCTcAAAAGCGTAT Adenomatous SEQ ID NO: 92 polyposis coli 591
GCCTtATGGAATTTG Adenomatous SEQ ID NO: 93 polyposis coli 608
GCTgTAGATGGTGC Adenomatous SEQ ID NO: 94 polyposis coli 617
GTTggcactcttacttaccGGAGCCAGAC Adenomatous SEQ ID NO: 95 polyposis
coli 620 CTTACttacCGGAGCCAGA Adenomatous SEQ ID NO: 96 polyposis
coli 621 ACTTaCCGGAGCCAG Adenomatous SEQ ID NO: 97 polyposis coli
624 AGCcaGACAAACACT Adenomatous SEQ ID NO: 98 polyposis coli 624
AGCCagacAAACACTTTA Adenomatous SEQ ID NO: 99 polyposis coli 626
ACAaacaCTTTAGCCAT Adenomatous SEQ ID NO: 100 polyposis coli 629
TTAGCcATTATTGAAA Adenomatous SEQ ID NO: 101 polyposis coli 635
GGAGgTGGGATATTA Adenomatous SEQ ID NO: 102 polyposis coli 638
ATATtACGGAATGTG Adenomatous SEQ ID NO: 103 polyposis coli 639
TTACGgAATGTGTCCA Adenomatous SEQ ID NO: 104 polyposis coli 657
AGAgaGAACAACTGT Adenomatous SEQ ID NO: 105 polyposis coli 659
TATTTCAG_I14E15_GCaaatcctaagagagAACAA Adenomatous SEQ ID NO: 106
CTGTC polyposis coli 660 AACTgtCTACAAACTT Adenomatous SEQ ID NO:
107 polyposis coli 665 TTAttACAACACTTA Adenomatous SEQ ID NO: 108
polyposis coli 668 CACttAAAATCTCAT Adenomatous SEQ ID NO: 109
polyposis coli 673 AGTttgacaatagtCAGTAATGCA Adenomatous SEQ ID NO:
110 polyposis coli 768 CACTTaTCAGAAACTT Adenomatous SEQ ID NO: 111
polyposis coli 769 TTATcAGAAACTTTT Adenomatous SEQ ID NO: 112
polyposis coli 770 TCAGAaACTTTTGACA Adenomatous SEQ ID NO: 113
polyposis coli 780 AGTCcCAAGGCATCT Adenomatous SEQ ID NO: 114
polyposis coli 792 AAGCaAAGTCTCTAT Adenomatous SEQ ID NO: 115
polyposis coli 792 AAGCAaaGTCTCTATGG Adenomatous SEQ ID NO: 116
polyposis coli 793 CAAAgTCTCTATGGT Adenomatous SEQ ID NO: 117
polyposis coli 798 GATTatGTTTTTGACA Adenomatous SEQ ID NO: 118
polyposis coli 802 GACACcaatcgacatGATGATAATA Adenomatous SEQ ID NO:
119 polyposis coli 805 CGACatGATGATAATA Adenomatous SEQ ID NO: 120
polyposis coli 811 TCAGacaaTTTTAATACT Adenomatous SEQ ID NO: 121
polyposis coli 825 TATtTGAATACTAC Adenomatous SEQ ID NO: 122
polyposis coli 827 AATAcTACAGTGTTA Adenomatous SEQ ID NO: 123
polyposis coli 830 GTGTTacccagctcctctTCATCAAGAG Adenomatous SEQ ID
NO: 124 polyposis coli 833 AGCTCcTCTTCATCAA Adenomatous SEQ ID NO:
125 polyposis coli 836 TCATcAAGAGGAAGC Adenomatous SEQ ID NO: 126
polyposis coli 848 AAAGAtaGAAGTTTGGA Adenomatous SEQ ID NO: 127
polyposis coli 848 AAAGatagaagTTTGGAGAGA Adenomatous SEQ ID NO: 128
polyposis coli 855 GAACgCGGAATTGGT Adenomatous SEQ ID NO: 129
polyposis coli 856 CGCGgaattGGTCTAGGCA Adenomatous SEQ ID NO: 130
polyposis coli 856 CGCGgAATTGGTCTA Adenomatous SEQ ID NO: 131
polyposis coli 879 CAGaTCTCCACCAC Adenomatous SEQ ID NO: 132
polyposis coli
902 GAAGAcagaAGTTCTGGGT Adenomatous SEQ ID NO: 133 polyposis coli
907 GGGTcTACCACTGAA Adenomatous SEQ ID NO: 134 polyposis coli 915
GTGACaGATGAGAGAA Adenomatous SEQ ID NO: 135 polyposis coli 929
CATACacatTCAAACACTT Adenomatous SEQ ID NO: 136 polyposis coli 930
ACACAttcaAACACTTACA Adenomatous SEQ ID NO: 137 polyposis coli 931
CATtCAAACACTTA Adenomatous SEQ ID NO: 138 polyposis coli 931
CATTcAAACACTTAC Adenomatous SEQ ID NO: 139 polyposis coli 933
AACacttACAATTTCAC Adenomatous SEQ ID NO: 140 polyposis coli 935
TACAatttcactAAGTCGGAAA Adenomatous SEQ ID NO: 141 polyposis coli
937 TTCActaaGTCGGAAAAT Adenomatous SEQ ID NO: 142 polyposis coli
939 AAGtcggAAAATTCAAA Adenomatous SEQ ID NO: 143 polyposis coli 946
ACATgTTCTATGCCT Adenomatous SEQ ID NO: 144 polyposis coli 954
TTAGaaTACAAGAGAT Adenomatous SEQ ID NO: 145 polyposis coli 961
AATgATAGTTTAAA Adenomatous SEQ ID NO: 146 polyposis coli 963
AGTTTaAATAGTGTCA Adenomatous SEQ ID NO: 147 polyposis coli 964
TTAaataGTGTCAGTAG Adenomatous SEQ ID NO: 148 polyposis coli 973
TATGgTAAAAGAGGT Adenomatous SEQ ID NO: 149 polyposis coli 974
GGTAAaAGAGGTCAAA Adenomatous SEQ ID NO: 150 polyposis coli 975
AAAAgaGGTCAAATGA Thyroid cancer SEQ ID NO: 151 992 AGTAAgTTTTGCAGTT
Thyroid cancer SEQ ID NO: 152 993 AAGttttgcagttaTGGTCAATAC
Adenomatous SEQ ID NO: 153 polyposis coli 999 CAAtacccagCCGACCTAGC
Adenomatous SEQ ID NO: 154 polyposis coli 1023 ACACcAATAAATTAT
Adenomatous SEQ ID NO: 155 polyposis coli 1030 AAAtATTCAGATGA
Adenomatous SEQ ID NO: 156 polyposis coli 1032 TCAGatgagCAGTTGAACT
Adenomatous SEQ ID NO: 157 polyposis coli 1033 GATGaGCAGTTGAAC
Adenomatous SEQ ID NO: 158 polyposis coli 1049 TGGGcAAGACCCAAA
Adenomatous SEQ ID NO: 159 polyposis coli 1054 CACAtaataGAAGATGAAA
Adenomatous SEQ ID NO: 160 polyposis coli 1055 ATAAtagaaGATGAAATAA
Adenomatous SEQ ID NO: 161 polyposis coli 1056 ATAGAaGATGAAATAA
Adenomatous SEQ ID NO: 162 polyposis coli 1060 ATAAAacaaaGTGAGCAAAG
Adenomatous SEQ ID NO: 163 polyposis coli 1061 AAAcaaaGTGAGCAAAG
Adenomatous SEQ ID NO: 164 polyposis coli 1061 AAACaaAGTGAGCAAA
Adenomatous SEQ ID NO: 165 polyposis coli 1062 CAAAgtgaGCAAAGACAA
Adenomatous SEQ ID NO: 166 polyposis coli 1065 CAAAGacAATCAAGGAA
Adenomatous SEQ ID NO: 167 polyposis coli 1067 CAAtcaaGGAATCAAAG
Adenomatous SEQ ID NO: 168 polyposis coli 1071 CAAAgtACAACTTATC
Adenomatous SEQ ID NO: 169 polyposis coli 1079 ACTGagAGCACTGATG
Adenomatous SEQ ID NO: 170 polyposis coli 1082 ACTGAtgATAAACACCT
Adenomatous SEQ ID NO: 171 polyposis coli 1084 GATaaacACCTCAAGTT
Adenomatous SEQ ID NO: 172 polyposis coli 1086 CACCtcAAGTTCCAAC
Adenomatous SEQ ID NO: 173 polyposis coli 1093 TTTGgACAGCAGGAA
Adenomatous SEQ ID NO: 174 polyposis coli 1098 TGTgtTTCTCCATAC
Adenomatous SEQ ID NO: 175 polyposis coli 1105 CGGgGAGCCAATGG
Thyroid cancer SEQ ID NO: 176 1110 TCAGAaACAAATCGAG Adenomatous SEQ
ID NO: 177 polyposis coli 1121 ATTAAtcaaAATGTAAGCC Adenomatous SEQ
ID NO: 178 polyposis coli 1131 CAAgAAGATGACTA Adenomatous SEQ ID
NO: 179 polyposis coli 1134 GACTAtGAAGATGATA Adenomatous SEQ ID NO:
180 polyposis coli 1137 GATgataaGCCTACCAAT Adenomatous SEQ ID NO:
181 polyposis coli 1146 CGTTAcTCTGAAGAAG Adenomatous SEQ ID NO: 182
polyposis coli 1154 GAAGaagaaGAGAGACCAA Adenomatous SEQ ID NO: 183
polyposis coli 1155 GAAGaagaGAGACCAACA Adenomatous SEQ ID NO: 184
polyposis coli 1156 GAAgagaGACCAACAAA Adenomatous SEQ ID NO: 185
polyposis coli 1168 GAAgagaaACGTCATGTG Adenomatous SEQ ID NO: 186
polyposis coli 1178 GATTAtagtttaAAATATGCCA Adenomatous SEQ ID NO:
187 polyposis coli 1181 TTAAaATATGCCACA Adenomatous SEQ ID NO: 188
polyposis coli 1184 GCCacagaTATTCCTTCA Adenomatous SEQ ID NO: 189
polyposis coli 1185 ACAgaTATTCCTTCA Adenomatous SEQ ID NO: 190
polyposis coli 1190 TCACAgAAACAGTCAT Adenomatous SEQ ID NO: 191
polyposis coli 1192 AAAcaGTCATTTTCA Adenomatous SEQ ID NO: 192
polyposis coli 1198 TCAaaGAGTTCATCT Adenomatous SEQ ID NO: 193
polyposis coli 1207 AAAAcCGAACATATG Adenomatous SEQ ID NO: 194
polyposis coli 1208 ACCgaacATATGTCTTC Adenomatous SEQ ID NO: 195
polyposis coli 1210 CATatGTCTTCAAGC Adenomatous SEQ ID NO: 196
polyposis coli 1233 CCAAGtTCTGCACAGA Adenomatous SEQ ID NO: 197
polyposis coli 1249 TGCAaaGTTTCTTCTA Adenomatous SEQ ID NO: 198
polyposis coli 1259 ATAcaGACTTATTGT Adenomatous SEQ ID NO: 199
polyposis coli 1260 CAGACttATTGTGTAGA Adenomatous SEQ ID NO: 200
polyposis coli 1268 CCAaTATGTTTTTC Adenomatous SEQ ID NO: 201
polyposis coli 1275 AGTtCATTATCATC Adenomatous SEQ ID NO: 202
polyposis coli 1294 CAGGAaGCAGATTCTG Adenomatous SEQ ID NO: 203
polyposis coli 1301 ACCCtGCAAATAGCA Adenomatous SEQ ID NO: 204
polyposis coli 1306 GAAAtaaaAGAAAAGATT Adenomatous SEQ ID NO: 205
polyposis coli 1307 ATAaAAGAAAAGAT Adenomatous SEQ ID NO: 206
polyposis coli 1308 AAAgaaaAGATTGGAAC Adenomatous SEQ ID NO: 207
polyposis coli 1308 AAAGAaaagaTTGGAACTAG Adenomatous SEQ ID NO: 208
polyposis coli 1318 GATCcTGTGAGCGAA Adenomatous SEQ ID NO: 209
polyposis coli 1320 GTGAGcGAAGTTCCAG Adenomatous SEQ ID NO: 210
polyposis coli 1323 GTTCcAGCAGTGTCA Adenomatous SEQ ID NO: 211
polyposis coli 1329 CACCctagaaccAAATCCAGCA Adenomatous SEQ ID NO:
212 polyposis coli 1336 AGACtgCAGGGTTCTA Adenomatous SEQ ID NO: 213
polyposis coli 1338 CAGgGTTCTAGTTT Adenomatous SEQ ID NO: 214
polyposis coli 1340 TCTAgTTTATCTTCA Adenomatous SEQ ID NO: 215
polyposis coli 1342 TTATcTTCAGAATCA Adenomatous SEQ ID NO: 216
polyposis coli 1352 GTTgAATTTTCTTC Adenomatous SEQ ID NO: 217
polyposis coli
1361 CCCTcCAAAAGTGGT Adenomatous SEQ ID NO: 218 polyposis coli 1364
AGTggtgCTCAGACACC Adenomatous SEQ ID NO: 219 polyposis coli 1371
AGTCCacCTGAACACTA Adenomatous SEQ ID NO: 220 polyposis coli 1372
CCACCtGAACACTATG Adenomatous SEQ ID NO: 221 polyposis coli 1376
TATGttCAGGAGACCC Adenomatous SEQ ID NO: 222 polyposis coli 1394
GATAgtTTTGAGAGTC Adenomatous SEQ ID NO: 223 polyposis coli 1401
ATTGCcAGCTCCGTTC Adenomatous SEQ ID NO: 224 polyposis coli 1415
AGTGGcATTATAAGCC Adenomatous SEQ ID NO: 225 polyposis coli 1426
AGCCcTGGACAAACC Adenomatous SEQ ID NO: 226 polyposis coli 1427
CCTGGaCAAACCATGC Adenomatous SEQ ID NO: 227 polyposis coli 1431
ATGCcACCAAGCAGA Adenomatous SEQ ID NO: 228 polyposis coli 1454
AAAAAtAAAGCACCTA Adenomatous SEQ ID NO: 229 polyposis coli 1461
GAAaAGAGAGAGAG Adenomatous SEQ ID NO: 230 polyposis coli 1463
AGAgagaGTGGACCTAA Adenomatous SEQ ID NO: 231 polyposis coli 1464
GAGAgTGGACCTAAG Adenomatous SEQ ID NO: 232 polyposis coli 1464
GAGAgtGGACCTAAGC Adenomatous SEQ ID NO: 233 polyposis coli 1464
GAGagTGGACCTAAG Adenomatous SEQ ID NO: 234 polyposis coli 1492
GCCaCGGAAAGTAC Adenomatous SEQ ID NO: 235 polyposis coli 1493
ACGGAaAGTACTCCAG Adenomatous SEQ ID NO: 236 polyposis coli 1497
CCAgATGGATTTTC Adenomatous SEQ ID NO: 237 polyposis coli 1503
TCAtccaGCCTGAGTGC Adenomatous SEQ ID NO: 238 polyposis coli 1522
TTAagaataaTGCCTCCAGT Adenomatous SEQ ID NO: 239 polyposis coli 1536
GAAACagAATCAGAGCA Adenomatous SEQ ID NO: 240 polyposis coli 1545
TCAAAtgaaaACCAAGAGAA Adenomatous SEQ ID NO: 241 polyposis coli 1547
GAAaACCAAGAGAA Adenomatous SEQ ID NO: 242 polyposis coli 1550
GAGAaagaGGCAGAAAAA Adenomatous SEQ ID NO: 243 polyposis coli 1577
GAATgtATTATTTCTG Adenomatous SEQ ID NO: 244 polyposis coli 1594
CCAGCcCAGACTGCTT Adenomatous SEQ ID NO: 245 polyposis coli 1596
CAGACtGCTTCAAAAT Adenomatous SEQ ID NO: 246 polyposis coli 1823
TTCAaTGATAAGCTC Adenomatous SEQ ID NO: 247 polyposis coli 1859
AATGAttctTTGAGTTCTC Adenomatous SEQ ID NO: 248 polyposis coli 1941
CCAGAcagaGGGGCAGCAA Desmoid SEQ ID NO: 249 tumours 1957
GAAaATACTCCAGT Adenomatous SEQ ID NO: 250 polyposis coli 1980
AACaATAAAGAAAA Adenomatous SEQ ID NO: 251 polyposis coli 1985
GAACCtATCAAAGAGA Adenomatous SEQ ID NO: 252 polyposis coli 1986
CCTaTCAAAGAGAC Adenomatous SEQ ID NO: 253 polyposis coli 1998
GAACcAAGTAAACCT Adenomatous SEQ ID NO: 254 polyposis coli 2044
AGCTCcGCAATGCCAA Adenomatous SEQ ID NO: 255 polyposis coli 2556
TCATCccttcctcGAGTAAGCAC Adenomatous SEQ ID NO: 256 polyposis coli
2643 CTAATttatCAAATGGCAC Adenomatous SEQ ID NO: 257 polyposis
coli
[0718] Bold letters indicate the codon. Undercase letters represent
the deletion. Where deletions extend beyond the coding region,
other positional information is provided. For example, the
abbreviation 5' UTR represents 5' untranslated region, and the
abbreviation E6I6 denotes exon 6/intron 6 boundary.
TABLE-US-00105 TABLE X SMALL INSERTIONS Codon Insertion Phenotype
157 T Adenomatous polyposis coli 170 AGAT Adenomatous polyposis
coli 172 T Adenomatous polyposis coli 199 G Adenomatous polyposis
coli 243 AG Adenomatous polyposis coli 266 T Adenomatous polyposis
coli 357 A Adenomatous polyposis coli 405 C Adenomatous polyposis
coli 413 T Adenomatous polyposis coli 416 A Adenomatous polyposis
coli 457 G Adenomatous polyposis coli 473 A Adenomatous polyposis
coli 503 ATTC Adenomatous polyposis coli 519 C Adenomatous
polyposis coli 528 A Adenomatous polyposis coli 561 A Adenomatous
polyposis coli 608 A Adenomatous polyposis coli 620 CT Adenomatous
polyposis coli 621 A Adenomatous polyposis coli 623 TTAC
Adenomatous polyposis coli 627 A Adenomatous polyposis coli 629 A
Adenomatous polyposis coli 636 GT Adenomatous polyposis coli 639 A
Adenomatous polyposis coli 704 T Adenomatous polyposis coli 740
ATGC Adenomatous polyposis coli 764 T Adenomatous polyposis coli
779 TT Adenomatous polyposis coli 807 AT Adenomatous polyposis coli
827 AT Adenomatous polyposis coli 831 A Adenomatous polyposis coli
841 CTTA Adenomatous polyposis coli 865 CT Adenomatous polyposis
coli 865 AT Adenomatous polyposis coli 900 TG Adenomatous polyposis
coli 921 G Adenomatous polyposis coli 927 A Adenomatous polyposis
coli 935 A Adenomatous polyposis coli 936 C Adenomatous polyposis
coli 975 A Adenomatous polyposis coli 985 T Adenomatous polyposis
coli 997 A Adenomatous polyposis coli 1010 TA Adenomatous polyposis
coli 1085 C Adenomatous polyposis coli 1085 AT Adenomatous
polyposis coli 1095 A Adenomatous polyposis coli 1100 GTTT
Adenomatous polyposis coli 1107 GGAG Adenomatous polyposis coli
1120 G Adenomatous polyposis coli 1166 A Adenomatous polyposis coli
1179 T Adenomatous polyposis coli 1187 A Adenomatous polyposis coli
1211 T Adenomatous polyposis coli 1256 A Adenomatous polyposis coli
1265 T Adenomatous polyposis coli 1267 GATA Adenomatous polyposis
coli 1268 T Adenomatous polyposis coli 1301 A Adenomatous polyposis
coli 1301 C Adenomatous polyposis coli 1323 A Adenomatous polyposis
coli 1342 T Adenomatous posis coli 1382 T Adenomatous polyposis
coli 1458 GTAG Adenomatous polyposis coli 1463 AG Adenomatous
polyposis coli 1488 T Adenomatous polyposis coli 1531 A Adenomatous
polyposis coli 1533 T Adenomatous polyposis coli 1554 A Adenomatous
polyposis coli 1555 A Adenomatous polyposis coli 1556 T Adenomatous
polyposis coli 1563 GACCT Adenomatous polyposis coli 1924 AA
Desmoid tumours indicates data missing or illegible when filed
TABLE-US-00106 TABLE XI SMALL INSERTIONS/DELETIONS Location/ codon
Deletion Insertion Phenotype 538 GAAGAcTTACAGCAGG gaa Adenomatous
SEQ ID NO: 258 polyposis coli 620 CTTACttaCCGGAGCCAG ct Adenomatous
SEQ ED NO: 259 polyposis coli 728 AATctcatGGCAAATAGG ttgcagctttaa
Adenomatous (SEQ ID NO: 260) (SEQ ID NO: polyposis coli 261) 971
GATGgtTATGGTAAAA taa Adenomatous SEQ ID NO: 262 polyposis coli
TABLE-US-00107 TABLE XII GROSS DELETIONS 2 kb including ex. 11
Adenomatous polyposis coli 3 kb I10E11 - 1.5 kb to I12E13 - 170 bp
Adenomatous polyposis coli 335 bp nt. 1409-1743 ex. 11-13
Adenomatous polyposis coli 6 kb incl. ex. 14 Adenomatous polyposis
coli 817 bp I13E14 - 679 to I13E14 + 138 Adenomatous polyposis coli
ex. 11-15M Adenomatous polyposis coli ex. 11-3'UTR Adenomatous
polyposis coli ex. 15A-ex. 15F Adenomatous polyposis coli ex. 4
Adenomatous polyposis coli ex. 7, 8 and 9 Adenomatous polyposis
coli ex. 8 to beyond ex. 15F Adenomatous polyposis coli ex. 8-ex.
15F Adenomatous polyposis coli ex. 9 Adenomatous polyposis coli
>10mb (del 5q22) Adenomatous polyposis coli
TABLE-US-00108 TABLE XIII GROSS INSERTIONS AND DUPLICATIONS
Description Phenotype Insertion of 14 bp nt. 3816 Adenomatous
polyposis coli Insertion of 22 bp nt. 4022 Adenomatous polyposis
coli Duplication of 43 bp cd. 1295 Adenomatous polyposis coli
Insertion of 337 bp of Alu I sequence cd. Desmoid tumours 1526
TABLE-US-00109 TABLE XIV COMPLEX REARRANGEMENTS (INCLUDING
INVERSIONS) A-T nt. 4893 Q1625H, Del C nt. 4897 cd. Adenomatous
polyposis coli 1627 Del 1099 bp I13E14 - 728 to E14I14 +
Adenomatous polyposis coli 156, ins 126 bp Del 1601 bp E14I14 + 27
to E14I14 + Adenomatous polyposis coli 1627, ins 180 bp Del 310 bp,
ins. 15 bp nt. 4394, cd 1464 Adenomatous polyposis coli Del A and T
cd. 1395 Adenomatous polyposis coli Del TC nt. 4145, Del TGT nt.
4148 Adenomatous polyposis coli Del. T, nt. 983, Del. 70 bp, nt.
985 Adenomatous polyposis coli Del. nt. 3892-3903, ins ATTT
Adenomatous polyposis coli
TABLE-US-00110 TABLE XV DIAGNOSTIC APPLICATIONS Cancer Type Marker
Application Reference Breast Her2/Neu Detection - Using methods
described herein, design D. Xie et al., J. Natl. polymorphism at
second primer such that after PCR, and Cancer Institute, 92, codon
655 digestion with restriction enzyme, a 5' 412 (2000) (GTC/valine
to overhang containing DNA sequence for K. S. Wilson et al.,
ATC/isoleucine codon 655 of Her2/Neu is generated. Am. J.
[Val(655)Ile]) Her2/Neu can be detected and Pathol., 161, 1171
quantified as a possible marker for (2002) breast cancer. Methods
described L. Newman, Cancer herein can detect both mutant allele
and Control, 9, 473 (2002) normal allele, even when mutant allele
is small fraction of total DNA. Herceptin therapy for breast cancer
is based upon screening for Her2. The earlier the mutant allele can
be detected, the faster therapy can be provided. Breast/Ovarian
Hypermethylation Methods described herein can be used M. Esteller
et al., of BRCA1 to differentiate between tumors New England Jnl
resulting from inherited BRCA1 Med., 344, 539 mutations and those
from non-inherited (2001) abnormal methylation of the gene Bladder
Microsatellite Methods described herein can be W. G. Bas et al.,
analysis of free applied to microsatellite analysis and Clinical
Cancer tumor DNA in FGFR3 mutation analysis for detection Res., 9,
257 (2003) Urine, Serum and of bladder cancer. Methods described M.
Utting et al., Plasma herein provide a non-invasive method Clinical
Cancer Res., for detection of bladder cancer. 8, 35 (2002) L. Mao,
D. Sidransky et al., Science, 271, 669 (1996) Lung Microsatellite
Methods described herein can be used T. Liloglou et al., analysis
of DNA to detect mutations in sputum samples, Cancer Research, 61,
from sputum and can markedly boost the accuracy of 1624, (2001)
preclinical lung cancer screening M. Tockman et al., Cancer
Control, 7, 19 (2000) Field et al., Cancer Research, 59, 2690
(1999) Cervical Analysis of HPV Methods described herein can be
used N. Munoz et al., New genotype to detect HPV genotype from a
cervical England Jnl Med., smear preparation. 348, 518 (2003) Head
and Neck Tumor specific Methods described herein can be used M.
Spafford et al. alterations in to detect any of 23 microsatellite
Clinical Cancer exfoliated oral markers, which are associated with
Research, 17, 607 mucosal cells Head and Neck Squamous Cell (2001)
(microsatellite Carcinoma (HNSCC). A. El-Naggar et al., J. markers)
Mol. Diag., 3, 164 (2001) Colorectal Screening for Methods
described herein can be used B. Ryan et al. mutation in K-ras2 to
detect K-ras 2 mutations, which can Gut, 52, 101 (2003) and APC
genes. be used as a prognostic indicator for colorectal cancer. APC
(see Example 5). Prostate GSTP1 Methods described herein can be
used P. Cairns et al. Clin. Hypermethylation to detect GSTP1
hypermethylation in Can. Res., 7, 2727 urine from patients with
prostate cancer; (2001) this can be a more accurate indicator than
PSA.
TABLE-US-00111 HIV Antiretroviral Screening Methods described
herein can be used J. Durant et al. resistance individuals for for
detection of mutations in the HIV The Lancet, 353, mutations in HIV
virus. Treatment outcomes are 2195 (1999) virus - e.g. 154V
improved in individuals receiving mutation or CCR5 anti-retroviral
therapy based upon .DELTA. 32 allele. resistance screening.
TABLE-US-00112 Cardiology Congestive Synergistic Methods described
herein can be used K. Small et al. New Heart Failure polymorphisms
of to genotype these loci and may help Eng. Jnl. Med., beta1 and
alpha2c identify people who are at a higher 347, 1135 (2002)
adrenergic risk of heart failure. receptors
Example 8
[0719] Single nucleotide polymorphisms (SNPs) represent the most
common form of sequence variation; three million common SNPs with a
population frequency of over 5% have been estimated to be present
in the human genome. A genetic map using these polymorphisms as a
guide is being developed
(http://research.marshfieldclinic.org/genetics/; internet address
as of Feb. 13, 2003).
[0720] The allele frequency varies from SNP to SNP; the allele
frequency for one SNP may be 50:50, while the allele frequency for
another SNP may be 90:10. The closer the allele frequency is to
50:50, the more likely any particular individual will be
heterozygous at that SNP. The SNP consortium provides allele
frequency information for some SNPs but not for others.
www.snp.chsl.org. The allele frequency for a particular SNP
provides valuable information as to the utility of that SNP for the
non-invasive prenatal screening method described in Example 5.
While all SNPs can be used, SNPs with allele frequencies closer to
50:50 are preferable.
[0721] Briefly, maternal blood contains fetal DNA. Maternal DNA can
be distinguished from fetal DNA by examining SNPs wherein the
mother is homozygous. For example, at SNP X, the maternal DNA may
be homozygous for guanine. If template DNA obtained from the plasma
of a pregnant female is heterozygous, as demonstrated by the
detection of signals corresponding to an adenine allele and an
guanine allele, the adenine allele can be used as a beacon for the
fetal DNA (see Example 5). The closer the allele frequency of a SNP
is to 50:50, the more likely there will be allele differences at a
particular SNP between the maternal DNA and the fetal DNA.
[0722] For example, if at SNP X the observed alleles are adenine
and guanine, and the SNP has an allele frequency of 90(A):10(G), it
is likely that both mother and father will be homozygous for
adenine at that particular SNP. Thus, both the maternal DNA and the
fetal DNA will be homozygous for adenine, and there is no distinct
signal for the fetal DNA. However, if at SNP X the allele frequency
is 50:50, and the mother is homozygous for adenine, the probability
is higher that the paternal DNA will contain a guanine allele at
SNP X.
[0723] Below, a method for determining the allele frequency for a
SNP is provided. Seven SNPs located on chromosome 13 were analyzed.
The method is applicable for any SNP including but not limited to
the SNPs on human chromosomes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,
12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X and Y.
Preparation of Template DNA
[0724] To determine the allele frequency of a particular SNP, DNA
was obtained from two hundred and fifty individuals after informed
consent had been granted. From each individual, a 9 ml blood sample
was collected into a sterile tube (Fischer Scientific, 9 ml EDTA
Vacuette tubes, catalog number NC9897284). The tubes were spun at
1000 rpm for ten minutes. The supernatant (the plasma) of each
sample was removed, and one milliliter of the remaining blood
sample, which is commonly referred to as the "buffy-coat" was
transferred to a new tube. One milliliter of 1.times.PBS was added
to each sample.
[0725] Template DNA was isolated using the QIAmp DNA Blood Midi Kit
supplied by QIAGEN (Catalog number 51183). The template DNA was
isolated as per instructions included in the kit. From each
individual, 0.76 .mu.g of DNA was pooled together, and the pooled
DNA was used in all subsequent reactions.
Design of Primers
[0726] SNP TSC0903430 was amplified using the following primer
set:
TABLE-US-00113 First primer: (SEQ ID NO: 279) 5'
GTCTTGCATGTAGAATTCTAGGGACGCTGCTTTTCGTC 3' Second primer: (SEQ ID
NO: 280) 5' CTCCTAGACATCGGGACTAGAATGTCCAC 3'
[0727] The first primer contained a recognition site for the
restriction enzyme EcoRI, and was designed to anneal eighty-two
bases from the locus of interest. The second primer contained the
recognition site for the restriction enzyme BsmF I.
[0728] SNP TSC0337961 was amplified using the following primer
set:
TABLE-US-00114 First primer: (SEQ D NO: 281) 5'
ACACAAGGCAGAGAATTCCAGTCCTGAGGGTGGGGGCC 3' Second primer: (SEQ ID
NO: 282) 5' CCGTGTTTTAACGGGACAAGCTGTTCTTC 3'
[0729] The first primer contained a recognition site for the
restriction enzyme EcoRI, and was designed to anneal ninety-two
bases from the locus of interest. The second primer contained the
recognition site for the restriction enzyme BsmF I.
[0730] SNP TSC0786441 was amplified using the following primer
set:
TABLE-US-00115 First primer: (SEQ ID NO: 283) 5'
GTAGCGGAGGTTGAATTCTATATGTTGTCTTGGACATT 3' Second primer: (SEQ ID
NO: 284) 5' CATCAGTAGAGTGGGACGAAAGTTCTGGC 3'
[0731] The first primer contained a recognition site for the
restriction enzyme EcoRI, and was designed to anneal one hundred
and four bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0732] SNP TSC1168303 was amplified using the following primer
set:
TABLE-US-00116 First primer: (SEQ ID NO: 285) 5'
ATCCACGCCGCAGAATTCGTATTCATGGGCATGTCAAA 3' Second primer: (SEQ ID
NO: 286) 5' CTTGGGACTATTGGGACCAGTGTTCAATC 3'
[0733] The first primer contained a recognition site for the
restriction enzyme EcoRI, and was designed to anneal sixty-four
bases from the locus of interest. The second primer contained the
recognition site for the restriction enzyme BsmF I.
[0734] SNP TSC0056188 was amplified using the following primer
set:
TABLE-US-00117 First primer: (SEQ ID NO: 287) 5'
CCAGAAAGCCGTGAATTCGTTAAGCCAACCTGACTCCA 3' Second primer: (SEQ ID
NO: 288) 5' TCGGGGTTAGTCGGGACATCCAGCAGCCC 3'
[0735] The first primer contained a recognition site for the
restriction enzyme EcoRI, and was designed to anneal eighty-two
bases from the locus of interest. The second primer contained the
recognition site for the restriction enzyme BsmF I.
[0736] SNP TSC0466177 was amplified using the following primer
s
TABLE-US-00118 First primer: (SEQ ID NO: 289) 5'
CGAAGGTAATGTGAATTCCAAAACTTAGTGCCACAATT 3' Second primer: (SEQ ID
NO: 290) 5' ATACCGCCCAACGGGACAGATCCATTGAC 3'
[0737] The first primer contained a recognition site for the
restriction enzyme EcoRI, and was designed to anneal ninety-two
bases from the locus of interest. The second primer contained the
recognition site for the restriction enzyme BsmF I.
[0738] SNP TSC0197424 was amplified using the following primer
set:
TABLE-US-00119 First primer: (SEQ ID NO: 291) 5'
AGAAACCTGTAAGAATTCGATTCCAAATTGTTTTTTGG 3' Second primer: (SEQ ID
NO: 292) 5' CGATCATAGGGGGGGACAGGAGAGAGCAC 3'
[0739] The first primer contained a recognition site for the
restriction enzyme EcoRI, and was designed to anneal one hundred
and four bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0740] The first primer was designed to anneal at various distances
from the locus of interest. The skilled artisan understands that
the annealing location of the first primer can be any distance from
the locus of interest including but not limited to 5-10, 11-15,
16-20, 21-25, 26-30, 31-35, 36-40, 41-45, 46-50, 51-55, 56-60,
61-65, 66-70, 71-75, 76-80, 81-85, 86-90, 91-95, 96-100, 101-105,
106-110, 111-115, 116-120, 121-125, 126-130, 131-140, 141-160,
161-180, 181-200, 201-220, 221-240, 241-260, 261-280, 281-300,
301-350, 351-400, 401-450, 451-500, 501-1000, 1001-2000, 2001-3000,
or greater than 3000.
[0741] All loci of interest were amplified from the template
genomic DNA using the polymerase chain reaction (PCR, U.S. Pat.
Nos. 4,683,195 and 4,683,202, incorporated herein by reference). In
this example, the loci of interest were amplified in separate
reaction tubes but they can also be amplified together in a single
PCR reaction. For increased specificity, a "hot-start" PCR was
used. PCR reactions were performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction can be optimized for each
locus of interest. In this example, 40 ng of template human genomic
DNA (a mixture of template DNA from 245 individuals) and 5 .mu.M of
each primer were used. Forty cycles of PCR were performed. The
following PCR conditions were used: [0742] (1) 95.degree. C. for 15
minutes and 15 seconds; [0743] (2) 37.degree. C. for 30 seconds;
[0744] (3) 95.degree. C. for 30 seconds; [0745] (4) 57.degree. C.
for 30 seconds; [0746] (5) 95.degree. C. for 30 seconds; [0747] (6)
64.degree. C. for 30 seconds; [0748] (7) 95.degree. C. for 30
seconds; [0749] (8) Repeat steps 6 and 7 thirty nine (39) times;
[0750] (9) 72.degree. C. for 5 minutes.
[0751] In the first cycle of PCR, the annealing temperature was
about the melting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[0752] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results.
Purification of Fragment of Interest
[0753] The PCR products were separated from the unused PCR
reagents. After the PCR reaction, 1/2 of the reaction volume for
SNP TSC0903430, SNP TSC0337961, and SNP TSC0786441 were mixed
together in a single reaction tube. One-half the reaction volumes
for SNPs TSC1168303, TSC0056188, TSC0466177, and TSC0197424 were
pooled together in a single reaction tube. The un-used primers, and
nucleotides were removed from the reaction by using Qiagen MinElute
PCR purification kits (Qiagen, Catalog Number 28004). The reactions
were performed following the manufacturer's instructions supplied
with the columns.
Restriction Enzyme Digestion of Isolated Fragments
[0754] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in eppendorf tubes following the instructions supplied
with the restriction enzyme.
Incorporation of Labeled Nucleotide
[0755] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[0756] As discussed in detail in Example 6, the sequence of both
alleles of a SNP can be determined with one labeled nucleotide in
the presence of the other unlabeled nucleotides. The following
components were added to each fill in reaction: 1 .mu.l of
fluorescently labeled ddGTP, 0.5 .mu.l of unlabeled ddNTPs (40
.mu.M), which contained all nucleotides except guanine, 2 .mu.l of
10.times. sequenase buffer, 0.25 .mu.l of Sequenase, and water as
needed for a 20 .mu.l reaction. The fill in reaction was performed
at 40.degree. C. for 10 min. Sequenase was the DNA polymerase used
in this example. However, any DNA polymerase can be used for a
fill-in reaction including but not limited to E. coli DNA
polymerase, Klenow fragment of E. coli DNA polymerase I, T7 DNA
polymerase, T4 DNA polymerase, Tag polymerase, Pfu DNA polymerase,
Vent DNA polymerase, polymerase from bacteriophage 29, and
REDTaq.TM. Genomic DNA polymerase. Non-fluorescently labeled ddNTP
was purchased from Fermentas Inc. (Hanover, Md.). An other labeling
reagents were obtained from Amersham (Thermo Sequenase Dye
Terminator Cycle Sequencing Core Kit, US 79565).
Detection of the Locus of Interest
[0757] The sample was loaded into a lane of a 36 cm 5% acrylamide
(urea) gel (BioWhittaker Molecular Applications, Long Ranger Run
Gel Packs, catalog number 50691). The sample was electrophoresed
into the gel at 3000 volts for 3 min. The gel was run for 3 hours
on a sequencing apparatus (Hoefer SQ3 Sequencer). The gel was
removed from the apparatus and scanned on the Typhoon 9400 Variable
Mode Imager. The incorporated labeled nucleotide was detected by
fluorescence.
[0758] Below, a schematic of the 5' overhang for SNP TSC0056188 is
reproduced (where R indicates the variable site). The entire
sequence is not shown, only a portion of the overhang.
TABLE-US-00120 5'CCA 3'GGT R T C C Overhang position 1 2 3 4
[0759] As discussed in detail in Example 6, one nucleotide labeled
with one chemical moiety can be used to determine the sequence of
the alleles of a locus of interest. The observed nucleotides for
TSC0056188 on the 5' sense strand (here depicted as the top strand)
are adenine and guanine. The third position in the overhang on the
antisense strand is cytosine, which is complementary to guanine. As
the variable site can be adenine or guanine, fluorescently labeled
ddGTP in the presence of unlabeled dCTP, dTTP, and dATP was used to
determine the sequence of both alleles. The fill-in reactions for
an individual homozygous for guanine, homozygous for adenine or
heterozygous are diagrammed below.
[0760] Homozygous Adenine:
TABLE-US-00121 5'CCA A A G* 3'GGT T T C C Overhang position 1 2 3
4
[0761] Homozygous Guanine:
TABLE-US-00122 5'CCA G* 3'GGT C T C C Overhang position 1 2 3 4
[0762] Heterozygous:
TABLE-US-00123 Allele 1 5'CCA G* 3'GGT C T C C Overhang position 1
2 3 4 Allele 2 5'CCA A A G* 3'GGT T T C C Overhang position 1 2 3
4
[0763] As seen in FIG. 14, two bands were detected for SNP
TSC0056188. The lower band corresponded to DNA molecules tilled in
with ddGTP at position one complementary to the overhang, which is
representative of the guanine allele. The higher band, separated by
a single base from the lower band, corresponded to DNA molecules
filled in with ddGTP at position 3 complementary to the overhang.
This band represented the adenine allele. The intensity of each
band was strong, indicating that each allele was well represented
in the population. SNP TSC0056188 is representative of a SNP with
high allele frequency.
[0764] Below, a schematic of the 5' overhang generated after
digestion with BsmF I for SNP TSC0337961 is reproduced (where R
indicates the variable site). The entire sequence is not shown,
only a portion of the overhang.
TABLE-US-00124 5' GCCA 3' CGGT R G C T Overhang position 1 2 3
4
[0765] The observed nucleotides for SNP TSC0337961 on the 5' sense
strand (here depicted as the top strand) are adenine and guanine.
The third position in the overhang on the antisense strand was
cytosine, which is complementary to guanine. As the variable site
can be adenine or guanine, fluorescently labeled ddGTP in the
presence of unlabeled dCTP, dTTP, and dATP was used to determine
the sequence of both alleles. The fill-in reactions for an
individual homozygous for guanine, homozygous for adenine or
heterozygous are diagrammed below.
[0766] Homozygous for Guanine:
TABLE-US-00125 5' GCCA G* 3' CGGT C G C T Overhang position 1 2 3
4
[0767] Homozygous for Adenine:
TABLE-US-00126 5' GCCA A C G* 3' CGGT T G C T Overhang position 1 2
3 4
[0768] Heterozygous
TABLE-US-00127 Allele 1 5' GCCA G* 3' CGGT C G C T Overhang
position 1 2 3 4 Allele 2 5' GCCA A C G* 3' CGGT T G C T Overhang
position 1 2 3 4
[0769] As seen in FIG. 14, one band migrating at the position of
the expected lower molecular weight band was observed. This band
represented the DNA molecules filled in with ddGTP at position one
complementary to the overhang, which represents the guanine allele.
No band corresponding to the DNA molecules filled in with ddGTP at
position 3 complementary to the overhang was detected. SNP
TSC0337961 is representative of a SNP that is not highly variable
within the population.
[0770] Of the seven SNPs analyzed, four of the SNPs (TSC1168303,
TSC0056188, TSC0466177, and TSC0197424 had high allele frequencies.
Two bands of high intensity were seen for each of the four SNPs,
indicating that both alleles were well represented in the
population.
[0771] However, it is not necessary that the SNPs have allele
frequencies of 50:50 to be useful. All SNPs provide useful
information. The methods described herein provide a rapid technique
for determining the allele frequency of a SNP, or any variable site
including but not limited to point mutations. Allele frequencies of
50:50, 51:49, 52:48, 53:47, 54:46, 55:45, 56:46, 57:43, 58:42,
59:41, 60:40, 61:39, 62:38, 63:37, 64:36, 65:35, 66:34, 67:33,
68:32, 69:31, 70:30, 71:29, 72:28, 73:27, 74:26, 75:25, 76:24,
77:23, 78:22, 79:21, 80:20, 81:19, 82:18, 83:17, 84:16, 85:15,
86:14, 87:13, 88:12, 89:11, 90:10, 91:9, 92:8, 93:7, 94:6, 95:5,
96:4, 97:3, 98:2, 99:1 and 100:0 can be useful.
[0772] Two bands were seen for SNP TSC0903430. One band, the lower
molecular weight band represented the DNA molecules filled in with
labeled ddGTP. A band of weaker intensity was seen for the
molecules filled in with labeled ddGTP at position 3 complementary
to the overhang, which represented the cytosine allele. SNP
TSC0903430 represents a SNP with low allele frequency variation. In
the population, the majority of individuals carry the guanine
allele, but the cytosine allele is still present.
[0773] One band of high intensity was seen for SNP TSC0337961 and
SNP TSC0786441. The band detected for both SNP TSC0337961 and SNP
TSC0786441 corresponded to the DNA molecules filled in with ddGTP
at position 1 complementary to the overhang. No signal was detected
from DNA molecules that would have been filled in at position 3
complementary to the overhang, which would have represented the
second allele. SNP TSC0337961 and SNP TSC0786441 represent SNPs
with little variability in the population.
[0774] As demonstrated in FIG. 14, the first primer used to amplify
each locus of interest can be designed to anneal at various
distances from the locus of interest. This allows multiple SNPs to
be analyzed in the same reaction. By designing the first primer to
anneal at specified distances from the loci of interest, any number
of loci of interest can be analyzed in a single reaction including
but not limited to 1-10, 11-20, 21-30, 31-40, 41-50, 51-60, 61-70,
71-80, 81-90, 91-100, 101-110, 111-120, 121-130, 131-140, 141-150,
151-160, 161-170, 171-180, 181-190, 191-200, 201-300, 301-400,
401-500, and greater than 500.
[0775] As discussed in Example 6, some type IIs restriction enzymes
display alternate cutting patterns. For example, the type IIS
restriction enzyme BsmF I typically cuts 10/14 from its binding
site; however, the enzyme also can cut 11/15 from the binding site.
To eliminate the effect of the alternate cut, the labeled
nucleotide used for the fill-in reaction should be chosen such that
it is not complementary to position 0 of the overhang generated by
the 11/15 cut (discussed in detail in Example 6). For instance, if
you label with ddGTP, the nucleotide preceding the variable site on
the strand that is filled in should not be a guanine.
[0776] The 11/15 overhang generated by BsmF I or SNP TSC0056188 is
depicted below, with the variable site in bold-typeface:
TABLE-US-00128 Allele 1 5'CC 3'GG T C T C Overhang position 0 1 2 3
Allele 2 5'CC 3'GG T T T C Overhang position 0 1 2 3
[0777] After the fill-in reaction with labeled ddGTP, unlabeled
dATP, dTTP, and dCTP, the following molecules were generated:
TABLE-US-00129 11/15 Allele 1 5'CC A G* 3'GG T C T C Overhang
position 0 1 2 3 11/15 Allele 2 5'CC A A A G* 3'GG T T T C Overhang
position 0 1 2 3
[0778] Two signals were seen; one band corresponded to molecules
filled in with ddGTP at position one of the overhang, and the other
band corresponded to the molecules filled in with ddGTP at position
3 complementary to the overhang. These are the same DNA molecules
generated after the fill-in reaction of the 10/14 overhang. Thus,
the two bands can be compared without any ambiguity from the
alternate cut. This method of labeling with a single nucleotide
eliminates any errors generated from the alternate cutting
properties of the enzymes.
[0779] The methods described herein is applicable to determining
the allele frequency of any SNP including but not limited to SNPs
on human chromosomes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, X and Y.
Example 9
[0780] Heterozygous SNPs, by definition, differ by one nucleotide.
At a heterozygous SNP, allele 1 and allele 2 may be present at a
ratio of 1:1. However, it is possible that DNA polymerases can
incorporate one nucleotide at a faster rate than other nucleotides,
and thus the observed ratio of a heterozygous SNP may differ from
the theoretically expected 1:1 ratio.
[0781] Below, methods are described that allow efficient and
accurate quantitation for the expected ratio of allele 1 to allele
2 at a heterozygous SNP.
Preparation of Template DNA
[0782] Template DNA was obtained from twenty-four individuals after
informed consent had been granted. From each individual, a 9 ml
blood sample was collected into a sterile tube (Fischer Scientific,
9 ml EDTA Vacuette tubes, catalog number NC9897284). The tubes were
spun at 1000 rpm for ten minutes without brake. The supernatant
(the plasma) of each sample was removed, and one milliliter of the
remaining blood sample, which is commonly referred to as the
"buffy-coat" was transferred to a new tube. One milliliter of
1.times.PBS was added to each sample.
[0783] Template DNA was isolated using the Q1 Amp DNA Blood Midi
Kit supplied by QIAGEN (Catalog number 51183). The template DNA was
isolated as per instructions included in the kit.
Design of Primers
[0784] SNP TSC0607185 was amplified using the following primer
set:
TABLE-US-00130 First primer: (SEQ ID NO: 40) 5'
ACTTGATTCCGTGAATTCGTTATCAATAAATCTTACAT 3' Second primer: (SEQ ID
NO: 41) 5' CAAGTTGGATCCGGGACCCAGGGCTAACC 3'
[0785] SNP TSC1130902 was amplified using the following primer
set:
TABLE-US-00131 First primer: (SEQ ID NO: 34) 5'
TCTAACCATTGCGAATTCAGGGCAAGGGGGGTGAGATC 3' Second primer: (SEQ ID
NO: 35) 5' TGACTTGGATCCGGGACAACGACTCATCC 3'
[0786] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI. The second
primer contained the recognition site for the restriction enzyme
BsmF I. The first primer was designed to anneal at various
distances from the locus of interest.
[0787] The first primer for SNP TSC0607185 was designed to anneal
ninety bases from the locus of interest. The first primer for SNP
TSC1130902 was designed to anneal sixty bases from the locus of
interest.
[0788] All loci of interest were amplified from the template
genomic DNA using the polymerase chain reaction (PCR, U.S. Pat.
Nos. 4,683,195 and 4,683,202, incorporated herein by reference). In
this example, the loci of interest were amplified in separate
reaction tubes but they could also be amplified together in a
single PCR reaction. For increased specificity, a "hot-start" PCR
was used. PCR reactions were performed using the HotStarTaq Master
Mix Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction can be optimized for each
locus of interest but in this example, 40 ng of template human
genomic DNA and 5 .mu.M of each primer were used. Forty cycles of
PCR were performed. The following PCR conditions were used: [0789]
(1) 95.degree. C. for 15 minutes and 15 seconds; [0790] (2)
37.degree. C. for 30 seconds; [0791] (3) 95.degree. C. for 30
seconds; [0792] (4) 57.degree. C. for 30 seconds; [0793] (5)
95.degree. C. for 30 seconds; [0794] (6) 64.degree. C. for 30
seconds; [0795] (7) 95.degree. C. for 30 seconds; [0796] (8) Repeat
steps 6 and 7 thirty nine (39) times; [0797] (9) 72.degree. C. for
5 minutes.
[0798] In the first cycle of PCR, the annealing temperature was
about the melting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[0799] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results.
Purification of Fragment of Interest
[0800] The PCR products were separated from the genomic template
DNA. One half of the PCR reaction was transferred to a well of a
Streptawell, transparent, High-Bind plate from Roche Diagnostics
GmbH (catalog number 1 645 692, as listed in Roche Molecular
Biochemicals, 2001 Biochemicals Catalog). The first primers
contained a 5' biotin tag so the PCR products bound to the
Streptavidin coated wells while the genomic template DNA did not.
The streptavidin binding reaction was performed using a Thermomixer
(Eppendorf) at 1000 rpm for 20 min. at 37.degree. C. Each well was
aspirated to remove unbound material, and washed three times with
1.times.PBS, with gentle mixing (Kandpal et al., Nucl. Acids Res.
18:1789-1795 (1990); Kaneoka et al., Biotechniques 10:30-34 (1991);
Green et al., Nucl. Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[0801] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in the Streptawells following the instructions supplied
with the restriction enzyme. After digestion, the wells were washed
three times with PBS to remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[0802] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[0803] As discussed in detail in Example 6, the sequence of both
alleles of a SNP can be determined by using one labeled nucleotide
in the presence of the other unlabeled nucleotides. The following
components were added to each fill in reaction: 1 .mu.l of
fluorescently labeled ddGTP, 0.5 .mu.l of unlabeled ddNTPs (40
.mu.M), which contained all nucleotides except guanine, 2 .mu.l of
10.times. sequenase buffer, 0.25 .mu.l of Sequenase, and water as
needed for a 20 .mu.l reaction. The fill in reaction was performed
at 40.degree. C. for 10 min. Non-fluorescently labeled ddNTP was
purchased from Fermentas Inc. (Hanover, Md.). All other labeling
reagents were obtained from Amersham (Thermo Sequenase Dye
Terminator Cycle Sequencing Core Kit, US 79565).
[0804] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme. Digestion was performed for 1 hour
at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[0805] The samples were loaded into a lane of a 36 cm 5% acrylamide
(urea) gel (BioWhittaker Molecular Applications, Long Ranger Run
Gel Packs, catalog number 50691). The samples were electrophoresed
into the gel at 3000 volts for 3 min. The gel was run for 3 hours
on a sequencing apparatus (Hoefer SQ3 Sequencer). The gel was
removed from the apparatus and scanned on the Typhoon 9400 Variable
Mode Imager. The incorporated labeled nucleotide was detected by
fluorescence. A box was drawn around each band and the intensity of
the band was calculated using the Typhoon 9400 Variable Mode Imager
software.
[0806] Below, a schematic of the 5' overhang for SNP TSC0607185 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang (where R indicates the variable
site).
TABLE-US-00132 C C T R TGTC 3' ACAG 5' 4 3 2 1 Overhang
position
[0807] The observed nucleotides at the variable site for TSC0607185
on the 5' sense strand (here depicted as the top strand) are
cytosine and thymidine (depicted here as R). In this case, the
second primer anneals from the locus of interest, which allows the
fill-in reaction to occur on the anti-sense strand (depicted here
as the bottom strand). The antisense strand will be filled in with
guanine or adenine.
[0808] The second position in the 5' overhang is thymidine, which
is complementary to adenine, and the third position in the overhang
corresponds to cytosine, which is complementary to guanine.
Fluorescently labeled ddGTP in the presence of unlabeled dCTP,
dTTP, and dATP was used to determine the sequence of both alleles.
After the fill-in reaction, the following DNA molecules were
generated:
TABLE-US-00133 C C T C TGTC 3' Allele 1 G* ACAG 5' 4 3 2 1 Overhang
position C C T T TGTC 3' Allele 1 G* A A ACAG 5' 4 3 2 1 Overhang
position
[0809] The overhang generated by BsmF I cutting at 11/15 from the
recognition site at TSC0607185 is depicted below:
TABLE-US-00134 C T R T GTC 3' 11/15 CAG 5' 3 2 1 0 Overhang
position
[0810] As labeled ddGTP is used for the fill-in reaction, no new
signal will be generated from the molecules cut 11/15 from the
recognition site. Position 0 complementary to the overhang was
filled in with unlabeled dATP. Only signals generated from
molecules filled in with labeled ddGTP at position 1 complementary
to the overhang or molecules filled in with labeled ddGTP at
position 3 complementary to the overhang were seen.
[0811] Five of the twenty-four individuals were heterozygous for
SNP TSC0607185. As shown in FIG. 15, two bands were detected. The
lower molecular weight band corresponded to DNA molecules filled in
with ddGTP at position 1 complementary to the overhang. The higher
molecular weight band corresponded to DNA molecules filled in with
ddGTP at position 3 complementary to the overhang.
[0812] The ratio of the two alleles was calculated for each of the
five heterozygous samples (see Table XVI). The average ratio of
allele 2 to allele 1 was 1.000 with a standard deviation of 0.044.
Thus, the allele ratio at SNP TSC0607185 was highly consistent. The
experimentally calculated allele ratio for a particular SNP is
hereinafter referred to as the "p" value of the SNP. Analysis of
SNP TSC0607185 consistently will provide an allele ratio of 1:1,
provided that the number of genomes analyzed is of sufficient
quantity that no error is generated from statistical sampling.
[0813] If the sample contained a low number of genomes, it is
statistically possible that the primers will anneal to one
chromosome over another chromosome. For example, if the sample
contains 40 genomes, which corresponds to a total of 40 chromosomes
of allele 1 and 40 chromosomes of allele 2, the primers may anneal
to 40 chromosomes of allele 1 but only 35 chromosome of allele 2.
This would cause allele 1 to be amplified preferentially to allele
2, which would alter the ratio of allele 1 to allele 2. This
problem is eliminated by having a sufficient number of genomes in
the sample.
[0814] SNP TSC0607185 represents a SNP where the difference in the
nucleotide at the variable site does not affect the PCR reaction,
or digestion with the restriction enzyme or the fill-in reaction.
The use of one nucleotide labeled with one fluorescent dye assures
that the bands for one allele can be accurately compared to the
bands for the second allele. There is no added complication of
having to compare between two different lanes, or having to correct
for the quantum coefficients of the dyes. Additionally, any effect
from the alternate cutting properties of the type IIS restriction
enzymes has been removed.
TABLE-US-00135 TABLE XVI Ratio of allele 2 to allele 1 at SNPs
TSC0607185 and TSC1130902. SNP TSC0607185 SNP TSC1130902 Sample
Allele 1 Allele 2 Allele2/Allele1 Allele 1 Allele 2 Allele2/Allele
1 1 2382 2313 0.971033 5877 4433 0.754296 2 1581 1533 0.969639 3652
2695 0.737952 3 1795 1879 1.046797 5416 3964 0.730059 4 1921 1855
0.965643 3493 2663 0.762382 5 1618 1701 1.051298 3894 2808 0.721109
Average 1.000882 0.74116 STD 0.044042 0.017018
[0815] Below, a schematic of the 5' overhang for SNP TSC1130902 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang (where R indicates the variable
site).
TABLE-US-00136 5' TTCAT 3' AAGTA R T C C Overhang position 1 2 3
4
[0816] The observed nucleotides for TSC1130902 on the 5' sense
strand (here depicted as the top strand) are adenine and guanine.
The second position in the overhang corresponds to a thymidine, and
the third position in the overhang corresponds to cytosine, which
is complementary to guanine. Fluorescently labeled ddGTP in the
presence of unlabeled dCTP, dTTP, and dATP was used to determine
the sequence of both alleles. After the fill-in reaction, the
following DNA molecules were generated:
TABLE-US-00137 Allele 1 5' TTCAT G* 3' AAGTA C T C C Overhang
position 1 2 3 4 Allele 2 5' TTCAT A A G* 3' AAGTA T T C C Overhang
position 1 2 3 4
[0817] As shown in FIG. 15, two bands were detected. The lower
molecular weight band corresponded to DNA molecules filled in with
labeled ddGTP at position 1 complementary to the overhang (the G
allele). The higher molecular weight band, separated by a single
base from the lower band, corresponded to DNA molecules filled in
with ddGTP at position 3 complementary to the overhang (the A
allele).
[0818] Five of the twenty-four individuals were heterozygous for
SNP TSC1130902. As seen in FIG. 15, the band corresponding to
allele 1 was more intense than the band corresponding to allele 2.
This was seen for each of the five individuals. The actual
intensity of the band corresponding to allele 1 varied from
individual to individual but it was always more intense than the
band corresponding to allele 2. For the five individuals, the
average ratio of allele 2 to allele 1 was 0.74116, with a standard
deviation of 0.017018.
[0819] Template DNA was prepared from five different individuals.
Separate PCR reactions, separate restriction enzyme digestions, and
separate fill-in reactions were performed. However, for each
template DNA, the ratio of allele 2 to allele 1 was about 0.75. The
"p" value for this SNP was highly consistent.
[0820] For example, for SNP TSC1130902, the "p" value was 0.75. Any
deviation from this value, provided the sample contains an adequate
number of genomes to remove statistical sampling errors, will
indicate that there is an abnormal copy number of chromosome 13. If
there is an additional copy of allele 2, the "p" value will be
higher than the expected 0.75. However, if there is an addition
copy of allele 1, the "p" value will be lower than the expected
0.75. With the "p" value quantitated for a particular SNP, that SNP
can be used to determine the presence or absence of a chromosomal
abnormality. An accurate "p" value measured for a single SNP will
be sufficient to detect the presence of a chromosomal
abnormality.
[0821] There are several possible explanations for why the ratio of
one allele to the other allele at some SNPs varies from the
theoretically expected ratio of 1:1. First, it is possible that the
DNA polymerase incorporates one nucleotide faster than the other
nucleotide. As the alleles are being amplified by PCR, even a
slight preference for one nucleotide over the other may cause
variation from the expected 1:1 ratio. This potential preference
for one nucleotide over the other is not seen during the fill-in
reaction because a single nucleotide labeled with one dye is
used.
[0822] It is also possible that the variable nucleotide at the SNP
site influences the rate of denaturation of the two alleles. If
allele 1 contains a guanine and allele 2 contains an adenine, the
difference between the strength of the bonds for these nucleotides
may affect the rate at which the DNA strands separate. Again, it is
important to mention that the alleles are being amplified by PCR so
very subtle differences can make a large impact on the final
result. It is also possible that the variable nucleotide at the SNP
site influences the rate at which the two strands anneal after
separation.
[0823] Alternatively, it is possible that the type IIS restriction
enzyme cuts one allele preferentially to the other allele. As
discussed in detail above, type IIS restriction enzymes cut at a
distance from the recognition site. It is possible that the
variable nucleotide at the SNP site influences the efficiency of
the restriction enzyme digestion. It is possible that at some SNPs
the restriction enzyme cuts one allele with an efficiency of 100%,
while it cuts the other allele with an efficiency of 90%.
[0824] However, the fact that the ratio of allele 1 to allele 2
deviates from the theoretically expected ratio of 1:1, does not
influence or reduce the utility of that SNP. As demonstrated above,
the "p" value for each SNP is consistent among different
individuals.
[0825] The "p" value for any SNP can be calculated by analyzing the
template DNA of any number of heterozygous individuals including
but not limited to 1-10, 11-20, 21-30, 31-40, 41-50, 51-60, 61-70,
71-80, 81-90, 91-100, 101-110, 111-120, 121-130, 131-140, 141-150,
151-160, 161-170, 171-180, 181-190, 191-200, 201-210, 211-220,
221-230, 231-240, 241-250, 251-260, 261-270, 271-280, 281-290,
291-300, and greater than 300.
[0826] The methods described herein allow the "p" value for any SNP
to be determined. It is possible that some SNPs will behave more
consistently than other SNPs. In the human genome, there are over 3
million SNPs; it is not possible to speculate on how each SNP will
behave. The "p" value for each SNP will have to be experimentally
determined. The methods described herein allow identification of
SNPs that have highly consistent, and reproducible "p" values.
Example 10
[0827] As discussed in Example 9, the ratio of one allele to the
other allele at a particular SNP may vary from the theoretically
expected ratio of 50:50. These SNPs can be used to detect the
presence of additional chromosomes provided that the ratio of one
allele to the other allele remains linear in individuals with
chromosomal disorders. For example, at SNP X if the percentage of
allele 1 to allele 2 is 75:25, the expected percentage of allele 1
to allele 2 for an individual with Down's syndrome must be properly
adjusted to reflect the variation from the expected percentage at
this SNP.
[0828] The percentage of allele 1 to allele 2 for SNP TSC0108992 on
chromosome 21 was calculated using template DNA from four normal
individuals and template DNA from an individual with Down's
syndrome. As demonstrated below, the percentage of one allele to
the other allele was consistent and remained linear in an
individual with Down's syndrome.
Preparation of Template DNA
[0829] DNA was obtained from four individuals with a normal genetic
karyotype and an individual identified as having an extra copy of
chromosome 21 (Down's syndrome). Informed consent was obtained from
all individuals. Informed consent also was obtained from the
parents of the individual with Down's syndrome.
[0830] From each individual, a 9 ml blood sample was collected into
a sterile tube (Fischer Scientific, 9 ml EDTA Vacuette tubes,
catalog number NC9897284). Template DNA was isolated using the
QIAmp DNA Blood Midi Kit supplied by QIAGEN (Catalog number 51183).
The template DNA was isolated as per instructions included in the
kit.
Design of Primers
[0831] SNP TSC0108992 was amplified using the following primer
set:
TABLE-US-00138 First primer: (SEQ ID NO: 293) 5'
CTACTGAGGGCTCGTAGATCCCAATTCCTTCCCAAGCT 3' Second primer: (SEQ ID
NO: 294) 5' AATCCTGCTTTAGGGACCATGCTGGTGGA 3'
[0832] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI. The second
primer contained the recognition site for the restriction enzyme
BsmF I.
[0833] SNP TSC0108992 was amplified from the template genomic DNA
using the polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195
and 4,683,202, incorporated herein by reference). For increased
specificity, a "hot-start" PCR was used. PCR reactions were
performed using the HotStarTaq Master Mix Kit supplied by QIAGEN
(catalog number 203443). The amount of template DNA and primer per
reaction can be optimized for each locus of interest. In this
example, 50 ng of template human genomic DNA and 5 .mu.M of each
primer were used. Thirty-eight cycles of PCR were performed. The
following PCR conditions were used: [0834] (1) 95.degree. C. for 15
minutes and 15 seconds; [0835] (2) 37.degree. C. for 30 seconds;
[0836] (3) 95.degree. C. for 30 seconds; [0837] (4) 57.degree. C.
for 30 seconds; [0838] (5) 95.degree. C. for 30 seconds; [0839] (6)
64.degree. C. for 30 seconds; [0840] (7) 95.degree. C. for 30
seconds; [0841] (8) Repeat steps 6 and 7 thirty-seven (37) times;
[0842] (9) 72.degree. C. for 5 minutes.
[0843] In the first cycle of PCR, the annealing temperature was
about the melting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[0844] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results.
Purification of Fragment of Interest
[0845] The PCR products were separated from the genomic template
DNA. Each PCR reaction was split into two samples and transferred
to two separate wells of a Streptawell, transparent, High-Bind
plate from Roche Diagnostics GmbH (catalog number 1 645 692, as
listed in Roche Molecular Biochemicals, 2001 Biochemicals Catalog).
For each PCR reaction, there were two replicates; each in a
separate well of a microtiter plate. The first primer contained a
5' biotin tag so the PCR products bound to the Streptavidin coated
wells while the genomic template DNA did not. The streptavidin
binding reaction was performed using a Thermomixer (Eppendorf) at
1000 rpm for 20 min. at 37.degree. C. Each well was aspirated to
remove unbound material, and washed three times with 1.times.PBS,
with gentle mixing (Kandpal et al., Nucl. Acids Res. 18:1789-1795
(1990); Kaneoka et al., Biotechniques 10:30-34 (1991); Green et
al., Nucl. Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[0846] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in the Streptawells following the instructions supplied
with the restriction enzyme. After digestion, the wells were washed
three times with 1.times.PBS to remove the cleaved fragments.
Incorporation, of Labeled Nucleotide
[0847] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[0848] As discussed in detail in Example 6, the sequence of both
alleles of a SNP can be determined with one labeled nucleotide in
the presence of the other unlabeled nucleotides. The following
components were added to each fill in reaction: 1 .mu.l of
fluorescently labeled ddTTP, 0.5 .mu.l of unlabeled ddNTPs (40
.mu.M), which contained all nucleotides except thymidine, 2 .mu.l
of 10.times. sequenase buffer, 0.25 .mu.l of Sequenase, and water
as needed for a 20 .mu.l reaction. The fill in reaction was
performed at 40.degree. C. for 10 min. Non-fluorescently labeled
ddNTP was purchased from Fermentas Inc. (Hanover, Md.). All other
labeling reagents were obtained from Amersham (Thermo Sequenase Dye
Terminator Cycle Sequencing Core Kit, US 79565).
[0849] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme. Digestion was performed for 1 hour
at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[0850] The samples were loaded into the lanes of a 36 cm 5%
acrylamide (urea) gel (BioWhittaker Molecular Applications, Long
Ranger Run Gel Packs, catalog number 50691). The samples were
electrophoresed into the gel at 3000 volts for 3 min. The gel was
run for 3 hours on a sequencing apparatus (Hoefer SQ3 Sequencer).
The gel was removed from the apparatus and scanned on the Typhoon
9400 Variable Mode Imager. The incorporated labeled nucleotide was
detected by fluorescence. A box was drawn around each band and the
intensity of the band was calculated using the Typhoon 9400
Variable Mode Imager software.
[0851] Below, a schematic of the 5' overhang for SNP TSC0108992 is
shown. The entire DNA sequence is not reproduced, only the portion
to demonstrate the overhang (where R indicates the variable
site).
TABLE-US-00139 GTCC 3' G A C R CAGG 5' 4 3 2 1 Overhang
Position
[0852] The observed nucleotides for SNP TSC0108992 are adenine and
thymidine on the sense strand (here depicted as the top strand).
Position 3 of the overhang corresponds to adenine, which is
complementary to thymidine. Labeled ddTTP was used in the presence
of unlabeled dATP, dCTP, and dGTP. After the fill-in reaction with
labeled ddTTP, the following DNA molecules were generated:
TABLE-US-00140 T* G A GTCC 3' Allele 1 G A C T CAGG 5' 4 3 2 1
Overhang Position T* GTCC 3' Allele 2 G A C A CAGG 5' 4 3 2 1
Overhang Position
[0853] There was no difficulty in comparing the values obtained
from allele 1 to allele 2 because one labeled nucleotide was used
for the fill-in reaction, and the fill-in reaction for both alleles
occurred in a single tube. The alternate cutting properties of BsmF
I would not influence this analysis because the 11/15 overhang
would be filled in just as the 10/14 overhang. Schematics of the
filled-in 11/15 overhangs are depicted below:
TABLE-US-00141 T* G A G TCC 3' 11/15 Allele 1 A C T C AGG 5' 3 2 1
0 Overhang Position T* G TCC 3' 11/15 Allele 2 A C A C AGG 5' 3 2 1
0 Overhang Position
[0854] As seen in FIG. 16, two bands were seen for each sample of
template DNA. The lower molecular weight band corresponded to the
DNA molecules filled in with ddTTP at position one complementary to
the overhang, and the higher molecular weight band corresponded to
DNA molecules filled in with ddTTP at position 3 complementary to
the overhang.
[0855] The percentage of allele 2 to allele 1 was highly
consistent. (see Table XVII). In addition, for any given
individual, the replicates of the PCR reaction showed similar
results (see Table XVII). The percentage of allele 2 to allele 1
was calculated by dividing the value of allele 2 by the sum of the
values for allele 1 and allele 2 (allele 2/(allele 1+allele 2)).
From four individuals, the average percentage of allele 2 to allele
1 was 0.4773 with a standard deviation of 0.0097. The percentage of
allele 2 to allele 1 on template DNA isolated from an individual
with Down's syndrome was 0.3086.
[0856] The theoretically expected percentage of allele 2 to allele
1 using template DNA from a normal individual is 0.50. However, the
experimentally determined percentage was 0.4773. The theoretically
expected percentage of allele 2 to allele 1 for an individual with
an extra copy of chromosome 21 is 0.33. The experimentally
determined percentage of allele 2 to allele 1 for SNP TSC0108992
was 0.3086.
[0857] The deviation from the theoretically expected percentage is
highly consistent and remains linear. The following formula
demonstrates that the percentage of allele 2 to allele 1 at SNP
TSC0108992 remains linear even on template DNA obtained from an
individual with an extra copy of chromosome 21:
0.47 0.50 = X 0.33 ##EQU00001## X = 0.3102 ##EQU00001.2##
[0858] If the percentage of allele 2 to allele 1 using template DNA
obtained from a normal individual is determined to be 0.47, then
the percentage of allele 2 to allele 1 using template DNA from an
individual with Down's syndrome should be 0.3102. The
experimentally determined ratio was 0.3086, with a standard
deviation of 0.00186. There is no difference between the predicted
percentage and the experimentally determined percentage of allele 2
to allele 1 on template DNA from an individual with Down's
syndrome.
[0859] The percentage of one allele to the other allele at a
particular SNP is highly consistent, reproducible, and linear. This
demonstrates that any SNP, regardless of the calculated percentage
for one allele to another, can be used to determine the presence or
absence of a chromosomal disorder.
TABLE-US-00142 TABLE XVII Percentage of Allele 2 to Allele 1 at SNP
TSC0108992. Sample Allele 2 Allele 1 2/(2 + 1) 1A 9568886 10578972
0.474933 1B 8330864 9221381 0.474632 2A 9801053 10345444 0.486489
2B 8970942 9603102 0.482983 3A 8676718 9211085 0.485063 3B 10847024
11420943 0.487113 4A 10512420 12227107 0.462297 4B 7883584 9055289
0.465414 MEAN 0.477366 STDEV 0.009654 DS 6797400 15138959 0.309869
DS 6025753 13586890 0.307238 MEAN 0.308554 STDEV 0.00186
Example 11
[0860] The percentage of allele 2 to allele 1 for a particular SNP
is highly consistent. Statistically significant deviation from the
experimentally determined ratio indicates the presence of a
chromosomal abnormality. Below, the percentage of allele 2 to
allele 1 at SNP TSC0108992 on chromosome 21 was calculated using
template DNA from a normal individual and template DNA from an
individual with Down's syndrome. Mixtures containing various
amounts of normal DNA and Down's syndrome DNA were prepared and
analyzed in a blind fashion.
Preparation of Template DNA
[0861] DNA was obtained from an individual with a normal genetic
karyotype and an individual identified as having an extra copy of
chromosome 21 (Down's syndrome). Informed consent was obtained from
both individuals. Informed consent also was obtained from the
parents of the individual with Down's syndrome.
[0862] From each individual, a 9 ml blood sample was collected into
a sterile tube (Fischer Scientific, 9 ml EDTA Vacuette tubes,
catalog number NC9897284). Template DNA was isolated using the
QIAmp DNA Blood Midi Kit supplied by QIAGEN (Catalog number 51183).
The template DNA was isolated as per instructions included in the
kit.
Mixtures of Template DNA
[0863] The template DNA from the individual with the normal
karyotype and the template DNA from the individual with an extra
copy of chromosome 21 were diluted to a concentration of 10
ng/.mu.l. Four mixtures of normal template DNA and Down's syndrome
template DNA were made in the following fashion:
[0864] Mixture 1: 32 .mu.l of Normal DNA+8 .mu.l of Down's syndrome
DNA
[0865] Mixture 2: 28 .mu.l of Normal DNA+12 .mu.l of Down's
syndrome DNA
[0866] Mixture 3: 20 .mu.l of Normal DNA+20 .mu.l of Down's
syndrome DNA
[0867] Mixture 4: 10 .mu.l of Normal DNA+30 .mu.l of Down's
syndrome DNA
[0868] Three separate PCR reactions were set up for the normal
template DNA and the template DNA from the individual with Down's
syndrome. Likewise, for each mixture, three separate PCR reactions
were set up.
Design of Primers
[0869] SNP TSC0108992 was amplified using the following primer
set:
TABLE-US-00143 First primer: (SEQ ID NO: 293) 5'
CTACTGAGGGCTCGTAGATCCCAATTCCTTCCCAAGCT 3' Second primer: (SEQ ID
NO: 294) 5' AATCCTGCTTTAGGGACCATGCTGGTGGA 3'
[0870] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI. The second
primer contained the recognition site for the restriction enzyme
BsmF I.
[0871] SNP TSC0108992 was amplified from the template genomic DNA
using the polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195
and 4,683,202, incorporated herein by reference).
[0872] For increased specificity, a "hot-start" PCR was used. PCR
reactions were performed using the HotStarTaq Master Mix Kit
supplied by QIAGEN (catalog number 203443). The amount of template
DNA and primer per reaction can be optimized for each locus of
interest but in this example, 50 ng of template human genomic DNA
and 5 .mu.M of each primer were used. Thirty-eight cycles of PCR
were performed. The following PCR conditions were used: [0873] (1)
95.degree. C. for 15 minutes and 15 seconds; [0874] (2) 37.degree.
C. for 30 seconds; [0875] (3) 95.degree. C. for 30 seconds; [0876]
(4) 57.degree. C. for 30 seconds; [0877] (5) 95.degree. C. for 30
seconds; [0878] (6) 64.degree. C. for 30 seconds; [0879] (7)
95.degree. C. for 30 seconds; [0880] (8) Repeat steps 6 and 7
thirty-seven (37) times; [0881] (9) 72.degree. C. for 5
minutes.
[0882] In the first cycle of PCR, the annealing temperature was
about the melting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[0883] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results.
Purification of Fragment of Interest
[0884] The PCR products were separated from the genomic template
DNA. Each PCR reaction was split into two samples and transferred
to two separate wells of a Streptawell, transparent, High-Bind
plate from Roche Diagnostics GmbH (catalog number 1 645 692, as
listed in Roche Molecular Biochemicals, 2001 Biochemicals Catalog).
For each PCR reaction, there were two replicates, each in a
separate well of a microtiter plate. The first primer contained a
5' biotin tag so the PCR products bound to the Streptavidin coated
wells while the genomic template DNA did not. The streptavidin
binding reaction was performed using a Thermomixer (Eppendorf) at
1000 rpm for 20 min. at 37.degree. C. Each well was aspirated to
remove unbound material, and washed three times with 1.times.PBS,
with gentle mixing (Kandpal et al., Nucl. Acids Res. 18:1789-1795
(1990); Kaneoka et al., Biotechniques 10:30-34 (1991); Green et al,
Nucl. Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[0885] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in the Streptawells following the instructions supplied
with the restriction enzyme. After digestion, the wells were washed
three times with 1.times.PBS to remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[0886] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[0887] As discussed in detail in Example 6, the sequence of both
alleles of a SNP can be determined with one labeled nucleotide in
the presence of the other unlabeled nucleotides. The following
components were added to each fill in reaction; 1 .mu.l of
fluorescently labeled ddTTP, 0.5 .mu.l of unlabeled ddNTPs (40
.mu.M), which contained all nucleotides except thymidine, 2 .mu.l
of 10.times. sequenase buffer, 0.25 .mu.l of Sequenase, and water
as needed for a 20 .mu.l reaction. The fill in reaction was
performed at 40.degree. C. for 10 min. Non-fluorescently labeled
ddTTP was purchased from Fermentas Inc. (Hanover, Md.). All other
labeling reagents were obtained from Amersham (Thermo Sequenase Dye
Terminator Cycle Sequencing Core Kit, US 79565).
[0888] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme. Digestion was performed for 1 hour
at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[0889] The samples were loaded into the lanes of a 36 cm 5%
acrylamide (urea) gel (BioWhittaker Molecular Applications, Long
Ranger Run Gel Packs, catalog number 50691). The samples were
electrophoresed into the gel at 3000 volts for 3 min. The gel was
run for 3 hours on a sequencing apparatus (Hoefer SQ3 Sequencer).
The gel was removed from the apparatus and scanned on the Typhoon
9400 Variable Mode Imager. The incorporated labeled nucleotide was
detected by fluorescence. A box was drawn around each band and the
intensity of the band was calculated using the Typhoon 9400
Variable Mode Imager software.
[0890] As seen in FIGS. 17 A-F, two bands were seen. The lower
molecular weight band corresponded to the DNA molecules filled in
with ddTTP at position one complementary to the overhang. The
higher molecular weight band corresponded to DNA molecules filled
in with ddTTP at position 3 complementary to the overhang.
[0891] The experiment was performed in a blind fashion. The tubes
were coded so that it was not known what tube corresponded to what
template DNA. After the gels were analyzed, each tube was grouped
into the following categories: normal template DNA, Down's syndrome
template DNA, 3:1 mixture of Down's syndrome template DNA to normal
DNA, 1:1 mixture of normal template DNA to Down's syndrome template
DNA, 1:2.3 mixture of Down's syndrome template DNA to normal
template DNA, and 1:4 mixture of Down's syndrome template DNA to
normal template DNA. Each replicate of each PCR reaction
successfully was grouped into the appropriate category, which
demonstrates that the method can be used to detect abnormal DNA
even if it represents only a small percentage of the total DNA.
[0892] The percentage of allele 2 to allele 1 for each replicate of
the three PCR reactions from normal template DNA are displayed in
Table XVIII (also see FIG. 17A). The average percentage of allele 2
to allele 1 was calculated by dividing the value of allele 2 by the
sum of the values for allele 1 and allele 2 (allele 2/(allele
1+allele 2)), which resulted in an average of 0.50025 with a
standard deviation of 0.002897. Thus, allele 1 and allele 2 were
present in a ratio of 50:50. While the intensity of the bands
varied from one PCR reaction to another (compare reaction 1 with
reaction 3), there was no difference in intensity within a PCR
reaction. Furthermore, the values obtained for the two replicates
of the PCR reactions were very similar. Most of the variation was
between PCR reactions and was likely attributable to pipetting
errors.
[0893] The percentage of allele 2 to allele 1 for each replicate of
the three PCR reactions from Down's syndrome template DNA are
displayed in Table XVIII (see FIG. 17B). The percentage of allele 2
to allele 1 was calculated by dividing the value of allele 2 by the
sum of the values for allele 1 and allele 2 (allele 2/allele
1+allele 2), which resulted in an average of 0.301314 with a
standard deviation of 0.012917. It is clear even upon analysis of
the gel by the naked eye that allele 1 is present in a higher copy
number than allele 2 (see FIG. 17B). Again, most of the variation
occurs between PCR reactions and not within the replicate of a PCR
reaction. The majority of the statistical variation likely resulted
from pipetting errors.
[0894] Analysis of a single SNP was sufficient to detect the
presence of the chromosomal abnormality. One SNP is sufficient
provided that the "p" value of the SNP is known and that there are
an adequate number of genomes so that statistical sampling error is
not introduced into the analysis. In this experiment, there were
approximately 5,000 genomes in each reaction.
[0895] The reactions that consisted of a mixture of Down's syndrome
template DNA to normal template DNA at a ratio of 3:1 were clearly
distinguishable from the normal template DNA, and the other
mixtures of DNA (see FIG. 17C). The calculated percentage of allele
2 to allele 1 was 0.319089 with a standard deviation of 0.004346
(see Table XVIII). Likewise, the reactions that consisted of a
mixture of Down's syndrome template DNA to normal template DNA at
ratios of 1:1, and 1:2.3 were distinguishable (see FIGS. 17D and
17E) and the values were statistically significant from all other
reactions (see Table XVIII).
[0896] As the amount of normal template DNA increased, the
percentage of allele 2 to allele 1 increased. With a mixture of
Down's syndrome template DNA to normal template DNA of 1:4, the
percentage of allele 2 to allele 1 was 0397642, with a standard
deviation of 0.001903 (see FIG. 17F). The difference between this
value and the value obtained from normal template DNA is
statistically significant. Thus, the methods described herein allow
the detection of a chromosomal abnormality even when the sample is
not a homogeneous sample of abnormal DNA.
[0897] As described above, the presence of a small fraction of DNA
with an abnormal copy number of chromosomes can be detected even
among a large presence of normal DNA. It was clear, even by the
naked eye, that as the amount of normal DNA increased and the
amount of Down's syndrome DNA decreased, the intensities of the
bands that corresponded to alleles 1 and 2 equalized.
[0898] The above example analyzed a SNP located on chromosome 21.
However, any SNP may be analyzed on any chromosome including but
not limited to human chromosomes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,
12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X, and Y and fetal
chromosomes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
17, 18, 19, 20, 21, 22, X, and Y. In addition, chromosomes from
non-human organisms can be analyzed using the above methods. Any
combination of chromosomes can be analyzed. In the above example,
an extra copy of a chromosome was detected. However, the same
methods can be used to detect monosomies.
TABLE-US-00144 TABLE XVIII Percentage of allele 2 to allele 1 at
SNP TSC0108992 using normal template DNA and Down's syndrome
template DNA. Allele 1 Allele 2 2/(2 + 1) Normal Template DNA 1A
2602115 2604525 0.500231 1B 2855846 2923860 0.505884 2A 1954765
1941929 0.498353 2B 2084476 2068106 0.498029 3A 2044147 2035719
0.498967 3B 1760291 1760543 0.500036 Mean 0.50025 STD 0.002897
Down's Syndrome 1A 4046926 1595581 0.282779 1B 4275341 1736260
0.288818 2A 2875698 1299509 0.311244 2B 2453615 1069635 0.303593 3A
3169338 1426643 0.310411 3B 3737440 1687286 0.311036 Mean 0.301314
STD 0.012917 3:1 (Down's:Normal) 1A 4067623 1980770 0.327487 1B
4058506 1899853 0.318855 2A 2315044 1085860 0.319286 2B 2686984
1243406 0.316357 3A 3880385 1790764 0.315767 3B 3718661 1724189
0.316781 Mean 0.319089 STD 0.004346 1:1 (Down's:Normal) 1A 3540255
1929840 0.352798 1B 4004085 2161443 0.350569 2A 2358009 1282132
0.35222 2B 2158132 1238377 0.364603 3A 3052330 1648677 0.350707 3B
3852682 2024012 0.344413 Mean 0.352552 STD 0.006618 1:2.3
(Down's:Normal) 1A 3109326 1942597 0.384526 1B 3392477 2118011
0.38436 2A 2824213 1758428 0.383715 2B 2069889 1249545 0.376433 3A
2335128 1433016 0.380298 3B 2916772 1797965 0.38135 Mean 0.38178
STD 0.003128 1:4 (Down's:Normal) 1A 3066524 2039636 0.399446 1B
3068284 2038770 0.399207 2A 2325477 1542526 0.398791 2B 2366122
1562218 0.397679 3A 2151205 1403120 0.394764 3B 2397046 1571360
0.395968 Mean 0.397642 STD 0.001903
Example 12
[0899] As discussed above in Example 9, the ratio for allele 1 to
allele 2 at a heterozygous SNP is constant. However, one factor
that can influence the ratio of allele 1 to allele 2 at a
heterozygous SNP is a low number of genomes. For example, if there
are 40 genomes, which means that there are a total of 40
chromosomes of allele 1 and 40 chromosomes of allele 2, it is
statistically possible that the primers may anneal to 40 of the
chromosomes with allele 1 but only 30 of the chromosomes with
allele 2. This will affect the ratio of allele 1 to allele 2, and
can erroneously influence the "p" value for a particular SNP.
[0900] Typically, whole genomic amplification, which employs
degenerate oligonucleotide PCR, is used to increase low quantities
of genomic DNA samples. Oligonucleotides of 8, 10, 12, or 14 bases
are used to amplify the genome. It is thought that the primers
anneal randomly throughout the genome, and will amplify a small
genomic DNA sample into hundreds-fold more DNA for genetic
analysis.
[0901] The methods described herein exploit the fact that typically
the whole genome is not of interest. Particular loci of interest
located on one chromosome, or on multiple chromosomes or on
chromosomes that represent the entire genome are selected for
analysis. Even if the loci of interest are located on chromosomes
for the entire genome, it is preferential to amplify the region of
those chromosomes that contain the loci of interest.
[0902] To overcome the limit of a low number of genomes, which is
often seen with fetal DNA obtained from the plasma of a pregnant
female, a multiplex method can be used to increase the number of
genomes. The method described below preferentially amplifies the
chromosome or chromosomes that contain the loci of interest.
Preparation of Template DNA
[0903] A 9 ml blood sample was collected into a sterile tube from a
human volunteer after informed consent had been granted. (Fischer
Scientific, 9 ml EDTA Vacuette tubes, catalog number NC9897284).
The tubes were spun at 1000 rpm for ten minutes. The supernatant
(the plasma) of each sample was removed, and one milliliter of the
remaining blood sample, which is commonly referred to as the
"buffy-coat" was transferred to a new tube. One milliliter of
1.times.PBS was added to each sample. Template DNA was isolated
using the QIAmp DNA Blood Midi Kit supplied by QIAGEN (Catalog
number 51183).
Design of Multiplex Primers
[0904] Primers were designed to anneal at various regions on
chromosome 21 to increase the copy number of the loci of interest
located on chromosome 21. The primers were 12 bases in length.
However, primers of any length can be used including but not
limited to 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,
18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34,
35, 36-45, 46-55, 56-65, 66-75, 76-85, 86-95, 96-105, 106-115,
116-125, and greater than 125 bases. Primers were designed to
anneal to both the sense strand and the antisense strand.
[0905] Nine SNPs located on chromosome 21 were analyzed:
TSC0397235, TSC0470003, TSC1649726, TSC1261039, TSC0310507,
TSC1650432, TSC1335008, TSC0128307, and TSC0259757. Any number of
SNPs can be analyzed including but not limited to 1-10, 11-20,
21-30, 31-40, 41-50, 51-60, 61-70, 71-80, 81-90, 91-100, 101-200,
201-300, 301-400, 401-500, 501-600, 601-700, 701-800, 801-900,
901-1000, 1001-2000, 2001-3000, 3001-4000, 4001-5000, 5001-6000,
6001-7000, 7001-8000, 8001-9000, 9001-10,000 and greater than
10,000.
[0906] For each of the 9 SNPs, a 12 base primer was designed to
anneal approximately 130 bases upstream of the loci of interest,
and a 12 base primer was designed to anneal approximately 130 bases
downstream of the loci of interest (herein referred to as the
multiplex primers). The multiplex primers can be designed to anneal
at any distance from the loci of interest including but not limited
to 10-20, 21-30, 31-40, 41-50, 51-60, 61-70, 71-80, 81-90, 91-100,
101-110, 111-120, 121-130, 131-140, 141-150, 151-160, 161-170,
171-180, 181-190, 191-200, 201-210, 211-220, 221-230, 231-240,
241-250, 251-260, 261-270, 271-280, 281-290, 291-300, 301-310,
311-320, 321-330, 331-340, 341-350, 351-360, 361-370, 371-380,
381-390, 391-400, 401-410, 411-420, 421-430, 431-440, 441-450,
451-460, 461-470, 471-480, 481-490, 491-500, 501-600, 601-700,
701-800, 801-900, 901-1000, 1001-2000, 2001-3000, 3001-4000,
4001-5000, and greater than 5000 bases. In addition, more than one
set of multiplex primers can be used for one SNP including but not
limited to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 10-20, 21-30, 31-40,
41-50, and greater than 50.
[0907] In addition, 91 sets of forward and reverse primers were
used to amplify other regions of chromosome 21, for a total of 100
sets of primers (200 primers in the reaction). These 91 primer sets
were used to demonstrate that a large number of primers can be used
in a single reaction without producing a large number of
non-specific bands. Any number of primers can be used in the
reaction including but not limited to 1-10, 11-20, 21-30, 31-40,
41-50, 51-60, 61-70, 71-80, 81-90, 91-100, 101-200, 201-300,
301-400, 401-500, 501-600, 601-700, 701-800, 801-900, 901-1000,
1001-2000, 2001-3000, 3001-4000, 4001-5000, 5001-6000, 6001-7000,
7001-8000, 8001-9000, 9001-10,000, 10,001-20,000, 20,001-30,000 and
greater than 30,000.
[0908] The multiplex primers were designed to have the same
nucleotides at the 3' end of the primer. In this case, the
multiplex primers ended in "AA," wherein A indicates adenine. The
primers were designed in this manner to minimize primer-dimer
formation. However, the primers can terminate in any nucleotides
including but not limited to adenine, guanine, cytosine, thymidine,
any combination of adenine and guanine, any combination of adenine
and cytosine, any combination of adenine and thymidine, any
combination of guanine and cytosine, any combination of guanine and
thymidine, or any combination of cytosine and thymidine. In
addition the multiplex primers can have 1, 2, 3, 4, 5, 6, 7, 8, 9,
10, or more than 10 of the same nucleotides at the 3' end.
[0909] The multiplex primers for SNP TSC0397235 were:
TABLE-US-00145 Forward Primer: 5' CAAGTGTCCTAA 3' (SEQ ID NO: 295)
Reverse primer: 5' CAGCTGCTAGAA 3' (SEQ ID NO: 296)
[0910] The multiplex primers for SNP TSC0470003 were:
TABLE-US-00146 Forward Primer: 5' GGTTGAGGGCAA 3' (SEQ ID NO: 297)
Reverse primer: 5' CACAGCGGGTAA 3' (SEQ ID NO: 298)
[0911] The multiplex primers for SNP TSC 1649726 were:
TABLE-US-00147 Forward Primer: 5' TTGACTTTTTAA 3' (SEQ ID NO: 299)
Reverse primer: 5' ACAGAATGGGAA 3' (SEQ ID NO: 300)
[0912] The multiplex primers for SNP TSC1261039 were:
TABLE-US-00148 Forward Primer: 5' TGCAGGTCACAA 3' (SEQ ID NO: 301)
Reverse primer: 5' TTCTTCTTATAA 3' (SEQ ID NO: 302)
[0913] The multiplex primers for SNP TSC0310507 were:
TABLE-US-00149 Forward Primer: 5' AGGACAACCTAA 3' (SEQ ID NO: 303)
Reverse primer: 5' TGGTGTTCAGAA 3' (SEQ ID NO: 304)
[0914] The multiplex primers for SNP TSC1650432 were:
TABLE-US-00150 Forward Primer: 5' TCAGCATATGAA 3' (SEQ ID NO: 305)
Reverse primer: 5' GTTGCCACACAA 3' (SEQ ID NO: 306)
[0915] The multiplex primers for SNP TSC1335008 were:
TABLE-US-00151 Forward Primer: 5' CCCAGCTAGCAA 3' (SEQ ID NO: 307)
Reverse primer: 5' GGGTCACTGTAA 3' (SEQ ID NO: 308)
[0916] The multiplex primers for SNP TSC0128307 were:
TABLE-US-00152 Forward Primer: 5' TTAAATACCCAA 3' (SEQ ID NO: 309)
Reverse primer: 5' TTAGGAGGTTAA 3' (SEQ ID NO: 310)
[0917] The multiplex primers for SNP TSC0259757 were:
TABLE-US-00153 Forward Primer: 5' ACACAGAATCAA 3' (SEQ ID NO: 311)
Reverse primer: 5' CGCTGAGGTCAA 3' (SEQ ID NO: 312)
[0918] Ninety-one (91) additional sets of primers, which annealed
to various regions along chromosome 21, were included in the
reaction:
TABLE-US-00154 Set 1: Forward Primer: 5' AAGTAGAGTCAA 3' (SEQ ID
NO: 313) Reverse primer: 5' CTTCCCATGGAA 3' (SEQ ID NO: 314) Set 2:
Forward Primer: 5' TTGGTTATTAAA 3' (SEQ ID NO: 315) Reverse primer:
5' CAACTTACTGAA 3' (SEQ ID NO: 316) Set 3: Forward Primer: 5'
CACTAAGTGAAA 3' (SEQ ID NO: 317) Reverse primer: 5' CTCACCTGCCAA 3'
(SEQ ID NO: 318) Set 4: Forward Primer: 5' ATGCATATATAA 3' (SEQ ID
NO: 319) Reverse primer: 5' AGAGATCAGCAA 3' (SEQ ID NO: 320) Set 5:
Forward Primer: 5' TATATTTTTCAA 3' (SEQ ID NO: 321) Reverse primer:
5' CAGAAAGCAGAA 3' (SEQ ID NO: 322) Set 6: Forward Primer: 5'
GTATTGGGTTAA 3' (SEQ ID NO: 323) Reverse primer: 5' CTGACCCAGGAA 3'
(SEQ ID NO: 324) Set 7: Forward Primer: 5' CAGTTTTCCCAA 3' (SEQ ID
NO: 325) Reverse primer: 5' AGGGCACAGGAA 3' (SEQ ID NO: 326) Set 8:
Forward Primer: 5' GTATCAGAGGAA 3' (SEQ ID NO: 327) Reverse primer:
5' GCATGAAAAGAA 3' (SEQ ID NO: 328) Set 9: Forward Primer: 5'
GATTTGACAGAA 3' (SEQ ID NO: 329) Reverse primer: 5' TACAGTTTACAA 3'
(SEQ ID NO: 330) Set 10: Forward Primer: 5' TGTGATTTTTAA 3' (SEQ ID
NO: 331) Reverse primer: 5' TTATGTTCTCAA 3' (SEQ ID NO: 332) Set
11: Forward Primer: 5' CAAGTACTTGAA 3' (SEQ ID NO: 333) Reverse
primer: 5' CTTGTGTGGCAA 3' (SEQ ID NO: 334) Set 12: Forward Primer:
5' AGACTTCTGCAA 3' (SEQ ID NO: 335) Reverse primer: 5' GTTGTCTTTCAA
3' (SEQ ID NO: 336) Set 13: Forward Primer: 5' GGGACACTCCAA 3' (SEQ
ID NO: 337) Reverse primer: 5' ATTATTATTCAA 3' (SEQ ID NO: 338) Set
14: Forward Primer: 5' ACATGATGACAA 3' (SEQ ID NO: 339) Reverse
primer: 5' TCAATTATAGAA 3' (SEQ ID NO: 340) Set 15: Forward Primer:
5' CTATGGGCTGAA 3' (SEQ ID NO: 341) Reverse primer: 5' TGTGTGCCTGAA
3' (SEQ ID NO: 342) Set 16: Forward Primer: 5' CCATTTGTTGAA 3' (SEQ
ID NO: 343) Reverse primer: 5' TCTCCATCAAAA 3' (SEQ ID NO: 344) Set
17: Forward Primer: 5' AATGCTGACAAA 3' (SEQ ID NO: 345) Reverse
primer: 5' TTTCATGTCCAA 3' (SEQ ID NO: 346) Set 18: Forward Primer:
5' GGCCTCTTGGAA 3' (SEQ ID NO: 347) Reverse primer: 5' TCATTTTTTGAA
3' (SEQ ID NO: 348) Set 19: Forward Primer: 5' GGACTACCATAA 3' (SEQ
ID NO: 349) Reverse primer: 5' AGTCACTCAGAA 3' (SEQ ID NO: 350) Set
20: Forward Primer: 5' CCTTGGCAGGAA 3' (SEQ ID NO: 351) Reverse
primer: 5' TTTCTGGTAGAA 3' (SEQ ID NO: 352) Set 21: Forward Primer:
5' CCCCCCCCCGAA 3' (SEQ ID NO: 353) Reverse primer: 5' GCCCAGGCAGAA
3' (SEQ ID NO: 354) Set 22: Forward Primer: 5' GAATGCGAAGAA 3' (SEQ
ID NO: 355) Reverse primer: 5' TTAGGTAGAGAA 3' (SEQ ID NO: 356) Set
23: Forward Primer: 5' TGCTTTGGTCAA 3' (SEQ ID NO: 357) Reverse
primer: 5' GCCCATTAATAA 3' (SEQ ID NO: 358) Set 24: Forward Primer:
5' TGAGATCTTTAA 3' (SEQ ID NO: 359) Reverse primer: 5' CAGTTTGTTCAA
3' (SEQ ID NO: 360) Set 25: Forward Primer: 5' GCTGGGCAAGAA 3' (SEQ
ID NO: 361) Reverse primer: 5' AGTCAAAGTCAA 3' (SEQ ID NO: 362) Set
26: Forward Primer: 5' TCTCTGCAGTAA 3' (SEQ ID NO: 363) Reverse
primer: 5' TGAATAACTTAA 3' (SEQ ID NO: 364) Set 27: Forward Primer:
5' CGGTTAGAAAAA 3' (SEQ ID NO: 365) Reverse primer: 5' CATCCCTTTCAA
3' (SEQ ID NO: 366) Set 28: Forward Primer: 5' TCTCTTTCTGAA 3' (SEQ
ID NO: 367) Reverse primer: 5' CTCAGATTGTAA 3' (SEQ ID NO: 368) Set
29: Forward Primer: 5' TTTGCACCAGAA 3' (SEQ ID NO: 369) Reverse
primer: 5' GGTTAACATGAA 3' (SEQ ID NO: 370) Set 30: Forward Primer:
5' ATTATCAACTAA 3' (SEQ ID NO: 371) Reverse primer: 5' GCCATTTTGTAA
3' (SEQ ID NO: 372) Set 31: Forward Primer: 5' GATCTAGATGAA 3' (SEQ
ID NO: 373) Reverse primer: 5' TTAATGTATTAA 3' (SEQ ID NO: 374) Set
32: Forward Primer: 5' CTAGGGAGACAA 3' (SEQ ID NO: 375) Reverse
primer: 5' TGGAGGAGACAA 3' (SEQ ID NO: 376) Set 33: Forward Primer:
5' CATCACATTTAA 3' (SEQ ID NO: 377) Reverse primer: 5' GGGGTCCTGCAA
3' (SEQ ID NO: 378) Set 34: Forward Primer: 5' CAGTTGTGCTAA 3' (SEQ
ID NO: 379) Reverse primer: 5' TCTGCAGCCTAA 3' (SEQ ID NO: 380) Set
35: Forward Primer: 5' GAGTCATTTAAA 3' (SEQ ID NO: 381) Reverse
primer: 5' TCTATGGATTAA 3' (SEQ ID NO: 382) Set 36: Forward Primer:
5' CAAAAAGTAGAA 3' (SEQ ID NO: 383) Reverse primer: 5' AATATACTCCAA
3' (SEQ ID NO: 384) Set 37: Forward Primer: 5' CGTCCAGCACAA 3' (SEQ
ID NO: 385) Reverse primer: 5' GGATGGTGAGAA 3' (SEQ ID NO: 386) Set
38: Forward Primer: 5' TCTCCTTTGTAA 3' (SEQ ID NO: 387) Reverse
primer: 5' TCGTTATTTCAA 3' (SEQ ID NO: 388) Set 39: Forward Primer:
5' GATTTTATAGAA 3' (SEQ ID NO: 389) Reverse primer: 5' AGACATAAGCAA
3' (SEQ ID NO: 390) Set 40: Forward Primer: 5' TTCACCTCACAA 3' (SEQ
ID NO: 391) Reverse primer: 5' GGATTGCTTGAA 3' (SEQ ID NO: 392) Set
41: Forward Primer: 5' ACTGCATGTGAA 3' (SEQ ID NO: 393) Reverse
primer: 5' TTTATCACAGAA 3' (SEQ ID NO: 394) Set 42: Forward Primer:
5' TCAGTAACACAA 3' (SEQ ID NO: 395) Reverse primer:
5' TACATCTTTGAA 3' (SEQ ID NO: 396) Set 43: Forward Primer: 5'
TTGTTTCAGTAA 3' (SEQ ID NO: 397) Reverse primer: 5' TATGAGCATCAA 3'
(SEQ ID NO: 398) Set 44: Forward Primer: 5' CTCAGCAGGCAA 3' (SEQ ID
NO: 399) Reverse primer: 5' ACCCCTGTATAA 3' (SEQ ID NO: 400) Set
45: Forward Primer: 5' TCTGCTCAGCAA 3' (SEQ ID NO: 401) Reverse
primer: 5' GTTCTTTTTTAA 3' (SEQ ID NO: 402) Set 46: Forward Primer:
5' GTGATAATCCAA 3' (SEQ ID NO: 403) Reverse primer: 5' GAGCCCTCAGAA
3' (SEQ ID NO: 404) Set 47: Forward Primer: 5' TTTATTGGTTAA 3' (SEQ
ID NO: 405) Reverse primer: 5' GGTACTGGGCAA 3' (SEQ ID NO: 406) Set
48: Forward Primer: 5' AGTGTTTTTCAA 3' (SEQ ID NO: 407) Reverse
primer: 5' TGTTATTGGTAA 3' (SEQ ID NO: 408) Set 49: Forward Primer:
5' GCGCATTCACAA 3' (SEQ ID NO: 409) Reverse primer: 5' AAACAAAAGCAA
3' (SEQ ID NO: 410) Set 50: Forward Primer: 5' TATATGATAGAA 3' (SEQ
ID NO: 411) Reverse primer: 5' TCCCAGTTCCAA 3' (SEQ ID NO: 412) Set
51: Forward Primer: 5' AAAGCCCATAAA 3' (SEQ ID NO: 413) Reverse
primer: 5' TGTCATCCACAA 3' (SEQ ID NO: 414) Set 52: Forward Primer:
5' TTGTGAATGCAA 3' (SEQ ID NO: 415) Reverse primer: 5' GTATTCATACAA
3' (SEQ ID NO: 416) Set 53: Forward Primer: 5' TGACATAGGGAA 3' (SEQ
ID NO: 417) Reverse primer: 5' AGCAAATTGCAA 3' (SEQ ID NO: 418) Set
54: Forward Primer: 5' AGTAGATGTTAA 3' (SEQ ID NO: 419) Reverse
primer: 5' AAAAGATAATAA 3' (SEQ ID NO: 420) Set 55: Forward Primer:
5' ACCTCATGGGAA 3' (SEQ ID NO: 421) Reverse primer: 5' TGGTCGACCTAA
3' (SEQ ID NO: 422) Set 56: Forward Primer: 5' TTTGCATGGTAA 3' (SEQ
ID NO: 423) Reverse primer: 5' GCGGCTGCCGAA 3' (SEQ ID NO: 424) Set
57: Forward Primer: 5' TCAGGAGTCTAA 3' (SEQ ID NO: 425) Reverse
primer: 5' GCCTACCAGGAA 3' (SEQ ID NO: 426) Set 58: Forward Primer:
5' ATCTTCTGTTAA 3' (SEQ ID NO: 427) Reverse primer: 5' AGGTAAGGACAA
3' (SEQ ID NO: 428) Set 59: Forward Primer: 5' TGCTTTGAGGAA 3' (SEQ
ID NO: 429) Reverse primer: 5' AACAGTTTTAAA 3' (SEQ ID NO: 430) Set
60: Forward Primer: 5' TTAAATGTTTAA 3' (SEQ ID NO: 431) Reverse
primer: 5' ATAGAAAATCAA 3' (SEQ ID NO: 432) Set 61: Forward Primer:
5' GTGTTGTGTTAA 3' (SEQ ID NO: 433) Reverse primer: 5' GAGGACCTCGAA
3' (SEQ ID NO: 434) Set 62: Forward Primer: 5' AGAGGCTGAGAA 3' (SEQ
ID NO: 435) Reverse primer: 5' GGTATTTATTAA 3' (SEQ ID NO: 436) Set
63: Forward Primer: 5' ATTTATCTGGAA 3' (SEQ ID NO: 437) Reverse
primer: 5' AGTGCAAACTAA 3' (SEQ ID NO: 438) Set 64: Forward Primer:
5' TGAACACCTTAA 3' (SEQ ID NO: 439) Reverse primer: 5' AATTTTTTCTAA
3' (SEQ ID NO: 440) Set 65: Forward Primer: 5' TTACTATTATAA 3' (SEQ
ID NO: 441) Reverse primer: 5' TGCTATAGTGAA 3' (SEQ ID NO: 442) Set
66: Forward Primer: 5' TGGACTATGGAA 3' (SEQ ID NO: 443) Reverse
primer: 5' CTGCAGTCCGAA 3' (SEQ ID NO: 444) Set 67: Forward Primer:
5' GCTACTGCCCAA 3' (SEQ ID NO: 445) Reverse primer: 5' TCACATGGTGAA
3' (SEQ ID NO: 446) Set 68: Forward Primer: 5' GTGGCTCTGGAA 3' (SEQ
ID NO: 447) Reverse primer: 5' GAATTCCATTAA 3' (SEQ ID NO: 448) Set
69: Forward Primer: 5' TGGGGTGTCCAA 3' (SEQ ID NO: 449) Reverse
primer: 5' GCAAGCTCCGAA 3' (SEQ ID NO: 450) Set 70: Forward Primer:
5' ATGTTTTTTCAA 3' (SEQ ID NO: 451) Reverse primer: 5' AGATCTGTTGAA
3' (SEQ ID NO: 452) Set 71: Forward Primer: 5' AAGTGCTGTGAA 3' (SEQ
ID NO: 453) Reverse primer: 5' ACTTTTTTGGAA 3' (SEQ ID NO: 454) Set
72: Forward Primer: 5' AATCGGCAGGAA 3' (SEQ ID NO: 455) Reverse
primer: 5' GGCATGTCACAA 3' (SEQ ID NO: 456) Set 73: Forward Primer:
5' AGGAAGAAAGAA 3' (SEQ ID NO: 457) Reverse primer: 5' CAGTTTCACCAA
3' (SEQ ID NO: 458) Set 74: Forward Primer: 5' CACAGAATTTAA 3' (SEQ
ID NO: 459) Reverse primer: 5' AAGAATAAGTAA 3' (SEQ ID NO: 460) Set
75: Forward Primer: 5' GGGATAGTACAA 3' (SEQ ID NO: 461) Reverse
primer: 5' TTCCCATGATAA 3' (SEQ ID NO: 462) Set 76: Forward Primer:
5' TGATTAGTTGAA 3' (SEQ ID NO: 463) Reverse primer: 5' GCATTCAGTGAA
3' (SEQ ID NO: 464) Set 77: Forward Primer: 5' AGGGAATATTAA 3' (SEQ
ID NO: 465) Reverse primer: 5' GACCTTAGGTAA 3' (SEQ ID NO: 466) Set
78: Forward Primer: 5' TTCTTTTCACAA 3' (SEQ ID NO: 467) Reverse
primer: 5' CCAAACTAAGAA 3' (SEQ ID NO: 468) Set 79: Forward Primer:
5' GTGCTCTTAGAA 3' (SEQ ID NO: 469) Reverse primer: 5' ATGAGTTTAGAA
3' (SEQ ID NO: 470) Set 80: Forward Primer: 5' ATGAGCATAGAA 3' (SEQ
ID NO: 471) Reverse primer: 5' GACAAATGAGAA 3' (SEQ ID NO: 472) Set
81: Forward Primer: 5' AAACCCAGAGAA 3' (SEQ ID NO: 473) Reverse
primer: 5' CCTCACACAGAA 3' (SEQ ID NO: 474) Set 82: Forward Primer:
5' CACACTGTGGAA 3' (SEQ ID NO: 475) Reverse primer: 5' CACTGTACCCAA
3' (SEQ ID NO: 476) Set 83: Forward Primer: 5' GTAGTATTTCAA 3' (SEQ
ID NO: 477) Reverse primer: 5' TGGATACACTAA 3' (SEQ ID NO: 478) Set
84: Forward Primer: 5' CCCATGATTCAA 3' (SEQ ID NO: 479)
Reverse primer: 5' TCATAGGAGGAA 3' (SEQ ID NO: 480) Set 85: Forward
Primer: 5' AGGAAAGAGAAA 3' (SEQ ID NO: 481) Reverse primer: 5'
ATATGGTGATAA 3' (SEQ ID NO: 482) Set 86: Forward Primer: 5'
GATGCCATCCAA 3' (SEQ ID NO: 483) Reverse primer: 5' ATACTATTTCAA 3'
(SEQ ID NO: 484) Set 87: Forward Primer: 5' GTGTGCATGGAA 3' (SEQ ID
NO: 485) Reverse primer: 5' AGGTGTTGAGAA 3' (SEQ ID NO: 486) Set
88: Forward Primer: 5' CAGCCTGGGCAA 3' (SEQ ID NO: 487) Reverse
primer: 5' GGAGCTCTACAA 3' (SEQ ID NO: 488) Set 89: Forward Primer:
5' AACTAAGGTTAA 3' (SEQ ID NO: 489) Reverse primer: 5' AACTTATGTTAA
3' (SEQ ID NO: 490) Set 90: Forward Primer: 5' ATCTCAACAGAA 3' (SEQ
ID NO: 491) Reverse primer: 5' TAACAATGTGAA 3' (SEQ ID NO: 492) Set
91: Forward Primer 5' AAGGATCAGGAA 3' (SEQ ID NO: 493) Reverse
primer: 5' CTCAAGTCTTAA 3' (SEQ ID NO: 494)
Multiplex PCR
[0919] Regions on chromosome 21 surrounding SNPs TSC0397235,
TSC0470003, TSC1649726, TSC1261039, TSC0310507, TSC1650432,
TSC1335008, TSC0128307, and TSC0259757 were amplified from the
template genomic DNA using the polymerase chain reaction (PCR, U.S.
Pat. Nos. 4,683,195 and 4,683,202, incorporated herein by
reference). This PCR reaction used primers that annealed
approximately 130 bases upstream and downstream of the loci of
interest. It was used to increases the number of copies of the loci
of interest to eliminate any errors that may result from a low
number of genomes.
[0920] For increased specificity, a "hot-start" PCR reaction was
used. PCR reactions were performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction can be optimized for each
locus of interest. In this example, 15 ng of template human genomic
DNA and 5 .mu.M of each primer were used.
[0921] Two microliters of each forward and reverse primer, at
concentrations of 5 mM were pooled into a single microcentrifuge
tube and mixed. Eight microliters of the primer mix was used in a
total PCR reaction volume of 40 .mu.l (1.5 .mu.l of template DNA,
10.5 .mu.l of sterile water, 8 .mu.l of primer mix, and 20 .mu.l of
HotStar Taq). Twenty-five cycles of PCR were performed. The
following PCR conditions were used: [0922] (1) 95.degree. C. for 15
minutes; [0923] (2) 95.degree. C. for 30 seconds; [0924] (3)
4.degree. C. for 30 seconds; [0925] (4) 37.degree. C. for 30
seconds; [0926] (5) Repeat steps 2-4 twenty-four (24) times; [0927]
(6) 72.degree. C. for 10 minutes.
[0928] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results.
[0929] In another embodiment, the loci of interest are amplified
using 6-base oligonucleotides, 7-base oligonucleotides, 8-base
oligonucleotides, 9-base oligonucleotides, 10-base
oligonucleotides, 11-base oligonucleotides, 12-base
oligonucleotides, 13-base oligonucleotides, 14-base
oligonucleotides, or greater than 14-base oligonucleotides. In a
preferred embodiment, 6-base oligonucleotides, 7-base
oligonucleotides, 8-base oligonucleotides, 9-base oligonucleotides,
10-base oligonucleotides, 11-base oligonucleotides, or 12-base
oligonucleotides are used to amplify the loci of interest. In
another embodiment, any number of oligonucleotides can be used
including but not limited to 1-5, 5-10, 10-15, 15-20, 20-25, 25-30,
30-35, 35-40, 40-45, 45-50, 50-100, 100-500, 500-1000, 1000-2000,
2000-4000, 4000-8000, 8000-10,000 or greater than 10,000. With a
small number of random oligos, the concentration of the oligos is
large enough to allow efficient amplification, and yet, the number
of oligos is small enough that it does not cause interference
between the oligos. This allows efficient amplification of the
genome.
[0930] In another embodiment, the upstream and downstream sequences
of the loci of interest are analyzed to identify a 6-base, 7-base,
8-base, 9-base, 10-base, 11-base, or 12-base sequence that is
present in the sequence upstream or downstream for each of the loci
of interest, which is then used to amplify the loci of interest. In
another embodiment, any number of 6-base oligonucleotides can be
used to amplify the loci of interest including but not limited to
1-10, 10-50, 50-100, 100-200, 200-500, or greater than 500.
[0931] In another embodiment, the number of loci of interest from a
small number of genomes can be increased by amplifying a limited
number of the loci of interest, followed by removal of the primers,
and amplification of the remaining loci of interest. All the loci
of interest do not have to be multiplexed in one reaction. Any
number of experimentally determined loci of interest can be
multiplexed in a single reaction including but not limited to 1-5,
5-10, 10-25, 25-50, 50-100, 100-200, 200-400, or greater than 400.
After increasing the number of copies of these loci of interest,
the sample can be passed through a column that allows the amplified
products to bind and the primers and unused dNTPs to be removed.
After eluting the bound products from the column, different loci of
interest can be amplified in a single reaction. This reduces the
amount of interaction between the primers.
[0932] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
Purification of Fragment of Interest
[0933] The excess primers and nucleotides were removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions were performed following the
manufacturer's instructions supplied with the columns. The DNA was
eluted in 100 .mu.l of sterile water.
PCR Reaction Two
[0934] SNP TSC0397235 was amplified using the following primer
set:
TABLE-US-00155 First Primer: (SEQ ID NO: 495) 5'
TTAGTCATCGCAGAATTCTACTTCTTTCTGAAGTGGGA 3' Second primer: (SEQ ID
NO: 496) 5' GGACAGCTCGATGGGACTAATGCATACTC 3'
[0935] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 103 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0936] SNP TSC0470003 was amplified using the following primer
set:
TABLE-US-00156 First Primer: (SEQ ID NO: 497) 5'
GTAGCCACTGGTGAATTCGTGCCATCGCAAAAGAATAA 3' Second primer: (SEQ ID
NO: 498) 5' ATTAGAATGATGGGGACCCCTGTCTTCCC 3'
[0937] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 80 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0938] SNP TSC1649726 was amplified using the following primer
set:
TABLE-US-00157 First Primer: (SEQ ID NO: 499) 5'
ACGCATAGGAAGGAATTCATTCTGACACGTGTGAGATA 3' Second primer: (SEQ ID
NO: 500) 5' GAAATTGACCACGGGACTGCACACTTTTC 3'
[0939] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 113 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0940] SNP TSC1261039 was amplified using the following primer
set:
TABLE-US-00158 First Primer: (SEQ ID NO: 501) 5'
CGGTAAATCGGAGAATTCAAGTTGAGGCATGCATCCAT 3' Second primer: (SEQ ID
NO: 502) 5' TCGGGGCTCAGCGGGACCACAGCCACTCC 3'
[0941] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 54 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0942] SNP TSC0310507 was amplified using the following primer
set:
TABLE-US-00159 First Primer: (SEQ ID NO: 503) 5'
TCTATGCACCACGAATTCAATATGTGTTCAAGGACATT 3' Second primer: (SEQ ID
NO: 504) 5' TGCTTAATCGGTGGGACTTGTAATTGTAC 3'
[0943] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 93 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0944] SNP TSC1650432 was amplified using the following primer
set:
TABLE-US-00160 First Primer: (SEQ ID NO: 505) 5'
CGCGTTGTATGCGAATTCCCTGGGGTATAAAGATAAGA 3' Second primer: (SEQ ID
NO: 506) 5' CTCACGGGAACTGGGACACCTGACCCTGC 3'
[0945] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 80 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0946] SNP TSC1335008 was amplified using the following primer
set:
TABLE-US-00161 First Primer: (SEQ ID NO: 507) 5'
GTCTTGCCGCTTGAATTCCCATAGAAGAATGCGCCAAA 3' Second primer: (SEQ ID
NO: 508) 5' TTGAGTAGTACAGGGACACACTAACAGAC 3'
[0947] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 94 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0948] SNP TSC0128307 was amplified using the following primer
set:
TABLE-US-00162 First Primer: (SEQ ID NO: 509) 5'
AATACTGTAGGTGAATTCTTGCCTAAGCATTTTCCCAG 3' Second primer: (SEQ ID
NO: 510) 5' GTGTTGACATTCGGGACTGTAATCTTGAC 3'
[0949] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 54 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0950] SNP TSC0259757 was amplified using the following primer
set:
TABLE-US-00163 First Primer: (SEQ ID NO: 511) 5'
TCTGTAGATTCGGAATTCTTTAGAGCCTGTGCGCTGAG 3' Second primer: (SEQ ID
NO: 512) 5' CGTACCAGTACAGGGACGCAAACTGAGAC 3'
[0951] The first primer contained a biotin tag at the 5' end and a
recognition site for the restriction enzyme EcoRI, and was designed
to anneal 100 bases from the locus of interest. The second primer
contained the recognition site for the restriction enzyme BsmF
I.
[0952] All loci of interest were amplified from the template
genomic DNA using the polymerase chain reaction (PCR, U.S. Pat.
Nos. 4,683,195 and 4,683,202, incorporated herein by reference). In
this example, the loci of interest were amplified in separate
reaction tubes but they can also be amplified together in a single
PCR reaction. For increased specificity, a "hot-start" PCR was
used. PCR reactions were performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443).
[0953] One microliter of the elutate from the multiplex reaction
(PCR product eluted from the MinElute column) was used as template
DNA for each PCR reaction. Each SNP was amplified in triplicate
when the multiplex sample was used as the template. As a control,
each SNP was amplified from 15 ng of the original template DNA (DNA
that did not undergo the multiplex reaction). The amount of
template DNA and primer per reaction can be optimized for each
locus of interest but in this example, 5 .mu.M of each primer was
used. Forty cycles of PCR were performed. The following PCR
conditions were used: [0954] (1) 95.degree. C. for 15 minutes and
15 seconds; [0955] (2) 37.degree. C. for 30 seconds; [0956] (3)
95.degree. C. for 30 seconds; [0957] (4) 57.degree. C. for 30
seconds; [0958] (5) 95.degree. C. for 30 seconds; [0959] (6)
64.degree. C. for 30 seconds; [0960] (7) 95.degree. C. for 30
seconds; [0961] (8) Repeat steps 6 and 7 thirty nine (39) times;
[0962] (9) 72.degree. C. for 5 minutes.
[0963] In the first cycle of PCR, the annealing temperature was
about the melting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[0964] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results.
Agarose Gel Analysis
[0965] Four microliters of a twenty microliter PCR reaction for
each SNP from the original template DNA was analyzed by agarose gel
electrophoresis (see FIG. 18A). Four microliters of a twenty
microliter PCR reaction for each SNP that was amplified from the
multiplexed template was analyzed on by agarose gel electrophoresis
(see FIG. 18B).
[0966] As seen in FIG. 18A, for 8/9 of the SNPs amplified from the
original template DNA, a single band of high intensity was seen
(lanes 1-3, and 5-9). The band migrated at the correct position for
each of the 8 SNPs. Amplification of TSC1261039 from the original
template DNA produced a band of high intensity, which migrated at
the correct position, and a faint band of lower molecular weight
(lane 4). Only two bands were seen, and the bands could clearly be
distinguished based on molecular weight. The PCR method described
herein allows clean amplification of the loci of interest from
genomic DNA without any concentration or enrichment of the loci of
interest.
[0967] As seen in FIG. 18B, the primers used to amplify SNPs
TSC0397235, TSC0470003, TSC0310507, and TSC0128307 from the
multiplexed template DNA produced a single band of high intensity,
which migrated at the correct position (lanes 1, 2, 5, and 8). No
additional bands were introduced despite the fact that the
multiplex reaction contained two hundred primers. While the
multiplex primers were 12 bases in length and likely annealed to
additional sequences other than those located on chromosome 21, the
products were not seen because the bands were not amplified in the
second PCR reaction. The second PCR reaction employed primers
specific for the loci of interest and used asymmetric
oligonucleotides and escalating annealing temperatures, which
allows specific amplification from the genome (see Example 1).
[0968] Amplification of TSC1649726 from the multiplex template DNA
produced one band of high intensity and two weaker bands, which
could clearly be distinguished based on molecular weight (see FIG.
18B, lane 3). Amplification of TSC1261039 from the multiplex
template DNA produced a high intensity band of the correct
molecular weight and a faint band of lower molecular weight (see
FIG. 18B, lane 4). The low molecular weight band was the same size
as the band seen from the amplification of TSC1261039 from the
original template DNA (compare FIG. 18A, lane 4 with FIG. 18B, lane
4). Thus, amplification of TSC1261039 on the multiplex template DNA
did not introduce any additional non-specific bands
[0969] Amplification of SNPs TSC1650432, TSC1335008, and TSC0259757
from the multiplex template DNA produced one band of high
intensity, which migrated at the correct position, and one weaker
band (lanes 6, 7, and 9). For SNPs TSC1650432 and TSC0259757, the
weaker band was of lower molecular weight, and clearly was
distinguishable from the band of interest (see FIG. 18B, lanes 6
and 9). For SNP TSC1335008, the weaker band was of slightly higher
molecular weight. However, the correct band can be identified by
comparing to the amplification products of TSC1335008 from the
original template DNA, (compare FIG. 18A, lane 7 and FIG. 18B, lane
7). The PCR conditions can also be optimized for TSC1335008. All 9
SNPs were amplified under the exact same conditions, which produced
clearly distinguishable bands for the amplified SNPs.
Purification of Fragment of Interest
[0970] The PCR products were separated from the genomic template
DNA. One half of the PCR reaction was transferred to a well of a
Streptawell, transparent, High-Bind plate from Roche Diagnostics
GmbH (catalog number 1 645 692, as listed in Roche Molecular
Biochemicals, 2001 Biochemicals Catalog). The first primers
contained a 5' biotin tag so the PCR products bound to the
Streptavidin coated wells while the genomic template DNA did not.
The streptavidin binding reaction was performed using a Thermomixer
(Eppendorf) at 1000 rpm for 20 min. at 37.degree. C. Each well was
aspirated to remove unbound material, and washed three times with
1.times.PBS, with gentle mixing (Kandpal et al., Nucl. Acids Res.
18:1789-1795 (1990); Kaneoka et al. Biotechniques 10:30-34 (1.991);
Green et al., Nucl. Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[0971] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in the Streptawells following the instructions supplied
with the restriction enzyme. After digestion, the wells were washed
three times with PBS to remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[0972] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNIP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[0973] As discussed in detail in Example 6, the sequence of both
alleles of a SNP can be determined by using one labeled nucleotide
in the presence of the other unlabeled nucleotides. The following
components were added to each fill in reaction: 1 .mu.l of
fluorescently labeled ddGTP, 0.5 .mu.l of unlabeled ddNTPs (40
.mu.M), which contained all nucleotides except guanine, 2 .mu.l of
10.times. sequenase buffer, 025 .mu.l of Sequenase, and water as
needed for a 20 .mu.l reaction. The fill in reaction was performed
at 40.degree. C. for 10 min. Non-fluorescently labeled ddNTP was
purchased from Fermentas Inc. (Hanover, Md.). All other labeling
reagents were obtained from Amersham (Thermo Sequenase Dye
Terminator Cycle Sequencing Core Kit, US 79565).
[0974] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments then were
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme. Digestion was performed for 1 hour
at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[0975] The samples were loaded into a lane of a 36 cm 5% acrylamide
(urea) gel (BioWhittaker Molecular Applications, Long Ranger Run
Gel Packs, catalog number 50691). The samples were electrophoresed
into the gel at 3000 volts for 3 min. The gel was run for 3 hours
on a sequencing apparatus (Hoefer SQ3 Sequencer). The gel was
removed from the apparatus and scanned on the Typhoon 9400 Variable
Mode Imager. The incorporated labeled nucleotide was detected by
fluorescence. A box was drawn around each band and the intensity of
the band was calculated using the ImageQuant software.
[0976] Below, a schematic of the 5' overhang for TSC0470003 after
digestion with BsmF I is depicted:
TABLE-US-00164 5' CTCT 3' GAGA R A C C Overhang position 1 2 3
4
[0977] The observed nucleotides for TSC0470003 are adenine and
guanine on the sense strand (herein depicted as the top strand).
The third position of the overhang corresponds to cytosine, which
is complementary to guanine. Labeled ddGTP was used in the presence
of unlabeled dATP, dCTP, and dTTP. Schematics of the DNA molecules
after the fill-in reaction are depicted below:
TABLE-US-00165 Allele 1 5' CTCT G* 3' GAGA C A C C Overhang
position 1 2 3 4 Allele 2 5' CTCT A T G* 3' GAGA T A C C Overhang
position 1 2 3 4
[0978] Two bands were seen; the lower molecular weight band
corresponded to the DNA molecules filled in with ddGTP at position
1 complementary to the overhang and the higher molecular weight
band corresponded to the DNA molecules filled in with ddGTP at
position 3 complementary to the overhang (see FIG. 19).
[0979] The percentage of allele 2 to allele 1 at TSC0470003 after
amplification from the original template DNA and the multiplexed
template DNA was calculated. The use of one fluorescently labeled
nucleotide to detect both alleles in a single reaction reduces the
amount of error that is introduced through pipetting reactions, and
the error that is introduced through the quantum coefficients of
different dyes.
[0980] For SNP TSC047003, the percentage of allele 2 to allele 1
was calculated by dividing the value of allele 2 by the sum of the
values for allele 2 and allele 1. The percentage of allele 2 to
allele 1 for TSC047003 on the original template DNA was calculated
to be 0.539 (see Table XIX). Three PCR reactions were performed for
each SNP on the multiplexed template DNA. The average percentage of
allele 2 to allele 1 for TSC047003 on the multiplexed DNA was 0.49
with a standard deviation of 0.0319 (see Table XIX). There was no
statistically significant difference between the percentage
obtained on the original template DNA and the multiplexed template
DNA.
[0981] For SNP TSC1261039, the percentage of allele 2 to allele 1
for TSC1261039 on the original template DNA was calculated to be
0.44 (see Table XIX). Three PCR reactions were performed for each
SNP on the multiplexed template DNA (see FIG. 19B). The average
percentage of allele 2 to allele 1 for TSC1261039 on the
multiplexed DNA was 0.468 with a standard deviation of 0.05683 (see
Table XIX). There was no statistically significant difference
between the percentages of allele 2 to allele 1 obtained on the
original template DNA and the multiplexed template DNA.
[0982] The variation seen in the percentage of allele 2 to allele 1
for TSC1261039 on the multiplexed template DNA was likely due to
pipetting reactions. The variation can be reduced by increasing the
number of replicates. With a large number of replicates, a
percentage can be obtained with minimum statistical variation.
[0983] Likewise, there was no statistical difference between the
percentage of allele 2 to allele 1 on the original template DNA and
on the multiplexed template DNA for SNPs TSC0310507 and TSC1335008
(see Table XIX, and FIGS. 19C and 19D). Thus, a multiplex reaction
can be used to increase the number of chromosomal regions
containing the loci of interest without affecting the percentage of
one allele to the other at the variable sites.
TABLE-US-00166 TABLE XIX Percentage of allele 2 to allele 1 at
various SNPs with and without multiplexing. Allele 1 Allele 2 2/(2
+ 1) TSC047003 IA 5535418 6487873 0.539608748 M1 4804358 4886716
0.504249168 M2 5549389 5958585 0.517778803 M3 8356275 7030245
0.45690936 Mean (M1-M3) 0.49297911 STDEV 0.031961429 TSC1261039 IA
3488765 2768066 0.442407027 M1 3603388 2573244 0.41660957 M2
4470423 5026872 0.529295131 M3 4306015 36694012 0.46008898 Mean
(M1-M3) 0.46866456 STDEV 0.056830136 TSC0310507 IA 2966511 2688190
0.475390299 M1 4084472 2963451 0.420471535 M2 4509891 4052892
0.47331481 M3 7173191 4642069 0.39288759 Mean (M1-M3) 0.428891312
STDEV 0.040869352 TSC1335008 IA 2311629 2553016 0.524810341 M1
794790 900879 0.531282343 M2 1261568 1780689 0.5853184 M3 1165156
1427840 0.550653 Mean (M1-M3) 0.555751248 STDEV 0.027376412
[0984] The methods described herein used two distinct amplification
reactions to amplify the loci of interest. In the first PCR
reaction, oligonucleotides were designed to anneal upstream and
downstream of the loci of interest. Unlike traditional genomic
amplification, these primers were not degenerate and annealed at a
specified distance from the loci of interest. However, due to the
length of the primers, it is likely that the primers annealed to
other regions of the genome. These primers were used to increase
the amount of DNA available for genetic analysis.
[0985] The second PCR reaction employs the methods described in
Examples 1-6. The primers are designed to amplify the loci of
interest, and the sequence is determined at the loci of interest.
The conditions of the second PCR reaction allowed specific
amplification of the loci of interest from the multiplexed template
DNA. If there were any non-specific products from the multiplex
reaction, they did not impede amplification of the loci of
interest. There was no statistical difference in the percentages of
allele 2 to allele 1 at the four SNPs analyzed, regardless of
whether the amplification was performed on original template DNA or
multiplexed template DNA.
[0986] The SNPs analyzed in this example were located on human
chromosome 21. However, the methods can be applied to non-human and
human DNA including but not limited to chromosomes 1, 2, 3, 4, 5,
6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X,
and Y. The multiplex methods can also be applied to analysis of
genetic mutations including but not limited to nucleotide
substitutions, insertions, deletions, and rearrangements.
[0987] The above methods can be used to increase the amount of DNA
available for genetic analysis whenever the starting template DNA
is limiting in quantity. For example, pre-malignant and
pre-invasive lesions with malignant cells usually constitute a
small fraction of the cells in the specimen, which reduces the
number of genetic analyses that can be performed. The methods
described herein can be used to increase the amounts of malignant
DNA available for genetic analysis. Also, the number of fetal
gnomes present in the maternal blood is often low; the methods
described herein can be used to increase the amount of fetal
DNA.
Example 13
[0988] Plasma isolated from blood of a pregnant female contains
both maternal template DNA and fetal template DNA. As discussed
earlier, the percentage of fetal. DNA in the maternal plasma varies
for each pregnant female. However, the percentage of fetal DNA can
be determined by analyzing SNPs wherein the maternal template DNA
is homozygous and the template DNA obtained from the plasma
displays a heterozygous pattern.
[0989] For example, assume SNP X can either be adenine or guanine,
and the maternal DNA for SNP X is homozygous for guanine. The
labeling method described in Example 6 can be used to determine the
sequence of the template DNA in the plasma sample. If the plasma
sample contains fetal DNA, which is heterozygous at SNP X, the
following DNA molecules are expected after digestion with the type
IIS restriction enzyme BsmF I, and the fill-in reaction with
labeled ddGTP, unlabeled dATP, dTTP, and dCTP.
TABLE-US-00167 Maternal Allele 1 5' GGGT G* 3'CCCA C T C A Maternal
Allele 2 5' GGGT G* 3'CCCA C T C A Fetal Allele 1 5' GGGT G* 3'CCCA
C T C A Fetal Allele 2 5' GGGT A A G* 3'CCCA T T C A
[0990] Two signals are seen; one signal corresponds to the DNA
molecules filled in with ddGTP at position one complementary to the
overhang and the second signal corresponds to the DNA molecules
filled in with ddGTP at position three complementary to the
overhang. However, the maternal DNA is homozygous for guanine,
which corresponds to the DNA molecules filled in at position one
complementary to the overhang. The signal from the DNA molecules
filled in with ddGTP at position three complementary to the
overhang corresponds to the adenine allele, which represents the
fetal DNA. This signal becomes a beacon for the fetal DNA, and can
used to measure the amount of fetal DNA present in the plasma
sample.
[0991] There is no difference in the amount of fetal DNA from one
chromosome to another. For instance, the percentage of fetal DNA in
any given individual from chromosome 1 is the same as the
percentage of fetal DNA from chromosome 2, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X and Y. Thus, the
allele ratio calculated for SNPs on one chromosome can be compared
to the allele ratio for the SNPs on another chromosome.
[0992] For example, the allele ratio for the SNPs on chromosome 1
should be equal to the allele ratio for the SNPs on chromosomes 2,
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
21, 22, X, and Y. However, if the fetus has a chromosomal
abnormality, including but not limited to a trisomy or monosomy,
the ratio for the chromosome that is present in an abnormal copy
number will differ from the ratio for the other chromosomes.
[0993] Blood from a pregnant female was collected after informed
consent had been obtained. The blood sample was used to demonstrate
that fetal DNA can be detected in the maternal plasma by analyzing
SNPs wherein the maternal DNA was homozygous, and the same SNP
displayed a heterozygous pattern from DNA obtained from the plasma
of a pregnant woman.
Preparation of Plasma from Whole Blood
[0994] Plasma was isolated from 4 tubes each containing 9 ml of
blood (Fischer Scientific, 9 ml EDTA Vacuette tubes, catalog number
NC9897284). The blood was obtained by venipuncture from a pregnant
female who had given informed consent. After collecting the blood,
formaldehyde (25 .mu.l/ml of blood) was added to each of the tubes.
The tubes were placed at 4.degree. C. until shipment. The tubes
were shipped via Federal Express in a foam container containing an
ice pack.
[0995] The blood was centrifuged at 1000 rpm for 10 minutes. The
brake on the centrifuge was not used. This centrifugation step was
repeated. The supernatant was transferred to a new tube and spun at
3,000 rpm for ten minutes. The brake on the centrifuge was not
used. The supernatant from each of the four tubes was pooled and
aliquoted into two tubes. The plasma was stored at -80.degree. C.
until the DNA was purified.
[0996] Template DNA was isolated using the QIAmp DNA Blood Midi Kit
supplied by QIAGEN (Catalog number 51183). The template DNA was
isolated as per instructions included in the kit. The template DNA
from the plasma was eluted in a final volume of 20 microliters.
Isolation of Maternal DNA
[0997] After the plasma was removed from the sample described
above, one milliliter of the remaining blood sample, which is
commonly referred to as the "buffy-coat," was transferred to a new
tube. One milliliter of 1.times.PBS was added to the sample.
Template DNA was isolated using the QIAmp DNA Blood Midi Kit
supplied by QIAGEN (Catalog number 51183).
Identification of Homozygous Maternal SNPs
[0998] Example 8 describes a method for identifying SNPs that are
highly variable within the population or for identifying
heterozygous SNPs for a given individual. The methods as described
in Example 8 were applied to the maternal template DNA to identify
SNPs on chromosome 13 wherein the maternal DNA was homozygous. Any
number of SNPs can be screened. The number of SNPs to be screened
is proportional to the number of heterozygous SNPs in the fetal DNA
that need to be analyzed.
[0999] As described in detail in Example 6, one labeled nucleotide
can be used to determine the sequence of both alleles at a
particular SNP. SNPs for which the sequence can be determined with
labeled ddGTP in the presence of unlabeled dATP, dTTP, and dCTP
were chosen for this example. However, SNPs for which the sequence
can be determined with labeled ddATP, ddCTP or ddTTP can also be
used. Additionally, the SNPs to be analyzed can be chosen such that
all are labeled with the same nucleotide or any combination of the
four nucleotides. For instance, if 400 SNPs are to be screened, 100
can be chosen such that the sequence is determined with labeled
ddATP, 100 can be chosen such that the sequence is determined with
labeled ddATP, 100 can be chosen such that the sequence is
determined with labeled ddGTP, and 100 can be chosen such that the
sequence is determined with labeled ddCTP, or any combination of
the four labeled nucleotides.
[1000] Twenty-nine SNPs wherein the maternal DNA was homozygous
were identified: TSC0052277, TSC1225391, TSC0289078, TSC1349804,
TSC0870209, TSC0194938, TSC0820373, TSC0902859, TSC0501510,
TSC1228234, TSC0082910, TSC0838335, TSC0818982, TSC0469204,
TSC1084457, TSC0466177, TSC1270598, TSC1002017, TSC1104200,
TSC0501389, TSC0039960, TSC0418134, TSC0603688, TSC0129188,
TSC1103570, TSC0813449, TSC0701940, TSC0087962, and TSC0660274.
Heterozygous SNPs will vary from individual to individual.
Design of Multiplex Primers
[1001] A low copy number of fetal genomes typically is present in
the maternal plasma. To increase the copy number of the loci of
interest located on chromosome 13, primers were designed to anneal
at approximately 130 bases upstream and 130 bases downstream of
each loci of interest. This was done to reduce statistical sampling
error that can occur when working with a low number of genomes,
which can influence the ratio of one allele to another (see Example
11). The primers were 12 bases in length. However, primers of any
length can be used including but, not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers were designed to anneal to both the sense strand and
the antisense strand.
[1002] The primers were designed to terminate at the 3' end in the
dinucleotide "AA" to reduce the formation of primer-dimers.
However, the primers can be designed to end in any of the four
nucleotides and in any combination of the four nucleotides.
[1003] The multiplex primers for SNPTSC0052277 were
TABLE-US-00168 Forward primer: 5' GACATGTTGGAA 3' (SEQ ID NO: 513)
Reverse primer: 5' ACTTCCAGTTAA 3' (SEQ ID NO: 514)
[1004] The multiplex primers for SNP TSC1225391 were:
TABLE-US-00169 Forward primer: 5' GTTTCCTGTTAA 3' (SEQ ID NO: 515)
Reverse primer 5' CGATGATGACAA 3' (SEQ ID NO: 516)
[1005] The multiplex primers for SNP TSC0289078 were:
TABLE-US-00170 Forward primer 5' GAGTAGAGACAA 3' (SEQ ID NO: 517)
Reverse primer 5' TCCCGGATACAA 3' (SEQ ID NO: 518)
[1006] The multiplex primes for SNP TSC1349804 were:
TABLE-US-00171 Forward primer: 5' CATCCTCTAGAA 3' (SEQ ID NO: 519)
Reverse primer: 5' TATTCCTGAGAA 3' (SEQ ID NO: 520)
[1007] The multiplex primers for SNP TSC0870209 were:
TABLE-US-00172 Forward primer: 5' AGTTTGTTTTAA 3' (SEQ ID NO: 521)
Reverse primer: 5' TATAAACGATAA 3' (SEQ ID NO: 522)
[1008] The multiplex primers for SNP TSC0194938 were:
TABLE-US-00173 Forward primer: 5' TTTGACCGATAA 3' (SEQ ID NO: 523)
Reverse primer: 5' TGACAGGACCAA 3' (SEQ ID NO: 524)
[1009] The multiplex primers for SNP TSC0820373 were:
TABLE-US-00174 Forward primer: 5' TTATTCATTCAA 3' (SEQ ID NO: 525)
Reverse primer: 5' AGTTTTTTCACAA 3' (SEQ ID NO: 526)
[1010] The multiplex primers for SNP TSC0902859 were:
TABLE-US-00175 Forward primer: 5' CACCTCCCTGAA 3' (SEQ ID NO: 527)
Reverse primer: 5' CCAGATTGAGAA 3' (SEQ ID NO: 528)
[1011] The multiplex primers for SNP TSC0501510 were:
TABLE-US-00176 Forward primer: 5' TGTGTCCACCAA 3' (SEQ ID NO: 529)
Reverse primer: 5' CTTCTATTCCAA 3' (SEQ ID NO: 530)
[1012] The multiplex primers for SNP TSC1228234 were:
TABLE-US-00177 Forward primer: 5' TCACAATAGGAA 3' (SEQ ID NO: 531)
Reverse primer 5' TACAAGTGAGAA 3' (SEQ ID NO: 532)
[1013] The multiplex primers for SNP TSC0082910 were:
TABLE-US-00178 Forward primer: 5' GAGTTTTCGTAA 3' (SEQ ID NO: 533)
Reverse primer: 5' GTGTGCCCCCAA 3' (SEQ ID NO: 534)
[1014] The multiplex primers for SNP TSC0838335 were:
TABLE-US-00179 Forward primer: 5' GCACCACTGCAA 3' (SEQ ID NO: 535)
Reverse primer: 5' GAACACAATGAA 3' (SEQ ID NO: 536)
[1015] The multiplex primers for SNP TSC0818982 were:
TABLE-US-00180 Forward primer: 5' TATCCTATTCAA 3' (SEQ ID NO: 537)
Reverse primer: 5' CAACCATTATAA 3' (SEQ ID NO: 538)
[1016] The multiplex primers for SNP TSC0469204 were:
TABLE-US-00181 Forward primer: 5' TATGCTTTACAA 3' (SEQ ID NO: 539)
Reverse primer: 5' TTTGTTTACCAA 3' (SEQ ID NO: 540)
[1017] The multiplex primers for SNP TSC1084457 were:
TABLE-US-00182 Forward primer: 5' AGGAAATTAGAA 3' (SEQ ID NO: 541)
Reverse primer: 5' TGTTAGACTTAA 3' (SEQ ID NO: 542)
[1018] The multiplex primers for SNP TSC0466177 were:
TABLE-US-00183 Forward primer: 5' TATTTGGAGGAA 3' (SEQ ID NO: 543)
Reverse primer: 5' GGCATTTGTCAA 3' (SEQ ID NO: 544)
[1019] The multiplex primers for SNP TSC1270598 were:
TABLE-US-00184 Forward primer: 5' ATACTCCAGGAA 3' (SEQ ID NO: 545)
Reverse primer: 5' CAGCCTGGACAA 3' (SEQ ID NO: 546)
[1020] The multiplex primers for SNP TSC1002017 were:
TABLE-US-00185 Forward primer: 5' CCATTGCAGTAA 3' (SEQ ID NO: 547)
Reverse primer: 5' AGGTTCTCATAA 3' (SEQ ID NO: 548)
[1021] The multiplex primers for SNP TSC1104200 were:
TABLE-US-00186 Forward primer: 5' TGTCATCATTAA 3' (SEQ ID NO: 549)
Reverse primer: 5' TGGTATTTGCAA 3' (SEQ ID NO: 550)
[1022] The multiplex primers for SNP TSC0501389 were:
TABLE-US-00187 Forward primer: 5' TAGGGTTTGTAA 3' (SEQ ID NO: 551)
Reverse primer: 5' CCCTAAGTAGAA 3' (SEQ ID NO: 552)
[1023] The multiplex primers for SNP TSC0039960 were:
TABLE-US-00188 Forward primer: 5' GTATTTCTTTAA 3' (SEQ ID NO: 553)
Reverse primer: 5' GAGTCTTCCCAA 3' (SEQ ID NO: 554)
[1024] The multiplex primers for SNP TSC0418134 were:
TABLE-US-00189 Forward primer: 5' CAGGTAGAGTAA 3' (SEQ ID NO: 555)
Reverse primer: 5' ATAGGATGTGAA 3' (SEQ ID NO: 556)
[1025] The multiplex primers for SNP TSC0603688 were:
TABLE-US-00190 Forward primer: 5' CAATGTGTATAA 3' (SEQ ID NO: 557)
Reverse primer: 5' AGAGGGCATCAA 3' (SEQ ID NO: 558)
[1026] The multiplex primers for SNP TSC0129188 were:
TABLE-US-00191 Forward primer: 5' CCAGTGGTCTAA 3' (SEQ ID NO: 559)
Reverse primer: 5' TAAACAATAGAA 3' (SEQ ID NO: 560)
[1027] The multiplex primers for SNP TSC1103570 were:
TABLE-US-00192 Forward primer: 5' GCACACTTTTAA 3' (SEQ ID NO: 561)
Reverse primer: 5' ATGGCTCTGCAA 3' (SEQ ID NO: 562)
[1028] The multiplex primers for SNP TSC0813449 were:
TABLE-US-00193 Forward primer: 5' GTCATCTTGTAA 3' (SEQ ID NO: 563)
Reverse primer: 5' TGCTTCATCTAA 3' (SEQ ID NO: 564)
[1029] The multiplex primers for SNP TSC0701940 were:
TABLE-US-00194 Forward primer: 5' AGAAAGGGGCAA 3' (SEQ ID NO: 565)
Reverse primer: 5' CTTTTCTTTCAA 3' (SEQ ID NO: 566)
[1030] The multiplex primers for SNP TSC0087962 were:
TABLE-US-00195 Forward primer: 5' CTACTCTCTCAA 3' (SEQ ID NO: 567)
Reverse primer: 5' ACAGCATTATAA 3' (SEQ ID NO: 568)
[1031] The multiplex primers for SNP TSC0660274 were:
TABLE-US-00196 Forward primer: 5' ACTGCTCTGGAA 3' (SEQ ID NO: 569)
Reverse primer: 5' GCAGAGGCACAA 3' (SEQ ID NO: 570)
[1032] Multiplex PCR
[1033] Regions on chromosome 13 surrounding the above-mentioned 29
SNPs were amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
used primers that annealed approximately 150 bases upstream and
downstream of each loci of interest. The fifty-eight primers were
mixed together and used in a single reaction to amplify the
template DNA. This reaction was done to increase the number of
copies of the loci of interest, which eliminates error generated
from a low number of genomes.
[1034] For increased specificity, a "hot-start" PCR reaction was
used. PCR reactions were performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction can be optimized for each
locus of interest. In this example, the 20 .mu.l of plasma template
DNA was used.
[1035] Two microliters of each forward and reverse primer, at
concentrations of 5 mM were pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix was used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.l of primer mix, and 25 .mu.l
of HotStar Taq. Twenty-five cycles of PCR were performed. The
following PCR conditions were used: [1036] (1) 95.degree. C. for 15
minutes; [1037] (2) 95.degree. C. for 30 second; [1038] (3)
4.degree. C. for 30 seconds; [1039] (4) 37.degree. C. for 30
seconds; [1040] (5) Repeat steps 2-4 twenty-four (24) times; [1041]
(6) 72.degree. C. for 10 minutes.
[1042] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results.
[1043] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
Purification of Fragment of Interest
[1044] The unused primers, and nucleotides were removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions were performed following the
manufacturer's instructions supplied with the columns. The DNA was
doted in 100 .mu.l of sterile water.
PCR Reaction Two
Design of Primers
[1045] SNPTSC0052277 was amplified using the following primer
set:
TABLE-US-00197 First primer: (SEQ ID NO: 571) 5'
CTCCGTGGTATGGAATTCCACTCAAATCTTCATTCAGA 3' Second primer: (SEQ ID
NO: 572) 5' ACGTCGGGTTACGGGACACCTGATTCCTC 3'
[1046] SNP TSC1225391 was amplified using the following primer
set:
TABLE-US-00198 First primer: (SEQ ID NO: 573) 5'
TACCATTGGTTTGAATTCTTGTTTCCTGTTAACCATGC 3' Second primer: (SEQ ID
NO: 574) 5' GCCGAGTTCTACGGGACAGAAAAGGGAGC 3'
[1047] SNP TSC0289078 was amplified using the following primer
set:
TABLE-US-00199 First primer: (SEQ ID NO: 575) 5'
TGCAGTGATTTCGAATTCGAGACAATGCTGCCCAGTCA 3' Second primer: (SEQ ID
NO: 576) 5' TCTAAATTCTCTGGGACCATTCCTTCAAC 3'
[1048] SNP TSC1349804 was amplified using the following primer
set:
TABLE-US-00200 First primer: (SEQ ID NO: 577) 5'
ACTAACAGCACTGAATTCCATGCTCTTGGACTTTCCAT 3' Second primer: (SEQ ID
NO: 578) 5' TCCCCTAACGTTGGGACACAGAATACTAC 3'
[1049] SNP TSC0870209 was amplified using the following primer
set:
TABLE-US-00201 First primer: (SEQ ID NO: 579) 5'
GTCGACGATGGCGAATTCCTGCCACTCATTCAGTTAGC 3' Second primer: (SEQ ID
NO: 580) 5' GAACGGCCCACAGGGACCTGGCATAACTC 3'
[1050] SNP TSC0194938 was amplified using the following primer
set:
TABLE-US-00202 First primer: (SEQ ID NO: 581) 5'
TCATGGTAGCAGGAATTCTGCTTTGACCGATAAGGAGA 3' Second primer: (SEQ ID
NO: 582) 5' ACTGTGGGATTCGGGACTGTCTACTACCC 3'
[1051] SNP TSC0820373 was amplified using the following primer
set:
TABLE-US-00203 First primer: (SEQ ID NO: 583) 5'
ACCTCTCGGCCGGAATTCGGAAAAGTGTACAGATCATT 3' Second primer: (SEQ ID
NO: 584) 5' GCCGGATACGAAGGGACGGCTCGTGACTC 3'
[1052] SNP TSC0902859 was amplified using the following primer
set:
TABLE-US-00204 First primer: (SEQ ID NO: 585) 5'
CCGTAGACTAAAGAATTCCCTGATGTCAGGCTGTCACC 3' Second primer: (SEQ ID
NO: 586) 5' ATCGGATCAGTCGGGACGGTGTCTTTGCC 3'
[1053] SNP TSC0501510 was amplified using the following primer
set:
TABLE-US-00205 First primer: (SEQ ID NO: 587) 5'
GCATAGGCGGGAGAATTCCCTGTGTCCACCAAAGTCGG 3' Second primer: (SEQ ID
NO: 588) 5' CCCACATAGGGCGGGACAAAGAGCTGAAC 3'
[1054] SNP TSC1228234 was amplified using the following primer
set:
TABLE-US-00206 First primer: (SEQ ID NO: 589) 5'
GGCTTGCCGAGCGAATTCTAGGAAAGATACGGAATCAA 3' Second primer: (SEQ ID
NO: 590) 5' TAACCCTCATACGGGACTTTCATGGAAGC 3'
[1055] SNP TSC0082910 was amplified using the following primer
set:
TABLE-US-00207 First primer: (SEQ ID NO: 591) 5'
ATGAGCACCCGGGAATTCTGATTGGAGTCTAGGCCAAA 3' Second primer: (SEQ ID
NO: 592) 5' TGCTCACCTTCTGGGACGTGGCTGGTCTC 3'
[1056] SNP TSC0838335 was amplified using the following primer
set:
TABLE-US-00208 First primer: (SEQ ID NO: 593) 5'
ACCGTCTGCCACGAATTCTGGAAAACATGCAGTCTGGT 3' Second primer: (SEQ ID
NO: 594) 5' TACACGGGAGGCGGGACAGGGTGATTAAC 3'
[1057] SNP TSC0818982 was amplified using the following primer
set:
TABLE-US-00209 First primer: (SEQ ID NO: 595) 5'
CTTAAAGCTAACGAATTCAGAGCTGTATGAAGATGCTT 3' Second primer: (SEQ ID
NO: 596) 5' AACGCTAAAGGGGGGACAACATAATTGGC 3'
[1058] SNP TSC0469204 was amplified using the following primer
set:
TABLE-US-00210 First primer: (SEQ ID NO: 597) 5'
TTGTAAGAACGAGAATTCTGCAACCTGTCTTTATTGAA 3' Second primer: (SEQ ID
NO: 598) 5' CTTCACCACTTTGGGACACTGAAGCCAAC 3'
[1059] SNP TSC1084457 was amplified using the following primer
set:
TABLE-US-00211 First primer: (SEQ ID NO: 599) 5'
AACCATTGATTTGAATTCGAAATGTCCACCAAAGTTCA 3' Second primer: (SEQ ID
NO: 600) 5' TGTCTAGTTCCAGGGACGCTGTTACTTAC 3'
[1060] SNP TSC0466177 was amplified using the following primer
set:
TABLE-US-00212 First primer: (SEQ ID NO: 601) 5'
CGAAGGTAATGTGAATTCTGCCACAATTAAGACTTGGA 3' Second primer: (SEQ ID
NO: 602) 5' ATACCGGTTTTCGGGACAGATCCATTGAC 3'
[1061] SNP TSC1270598 was amplified using the following primer
set:
TABLE-US-00213 First primer: (SEQ ID NO: 603) 5'
CCTGAAATCCACGAATTCCACCCTGGCCTCCCAGTGCA 3' Second primer: (SEQ ID
NO: 604) 5' TAGATGGTAGGTGGGACAGGACTGGCTTC 3'
[1062] SNP TSC1002017 was amplified using the following primer
set:
TABLE-US-00214 First primer: (SEQ ID NO: 605) 5'
GCATATCTTAGCGAATTCCTGTGACTAATACAGAGTGC 3' Second primer: (SEQ ID
NO: 606) 5' CCAAATATGGTAGGGACGTGTGAACACTC 3'
[1063] SNP TSC1104200 was amplified using the following primer
set:
TABLE-US-00215 First primer: (SEQ ID NO: 607) 5'
TGCCGCTACAGGGAATTCATATGGCAGATATTCCTGAA 3' Second primer: (SEQ ID
NO: 608) 5' ACGTTGCGGACCGGGACTTCCACAGAGCC 3'
[1064] SNP TSC0501389 was amplified using the following primer
set:
TABLE-US-00216 First primer: (SEQ ID NO: 609) 5'
CTTCGCCCAATGGAATTCGGTACAGGGGTATGCCTTAT 3' Second primer: (SEQ ID
NO: 610) 5' TGCACTTCTGCCGGGACCAGAGGAGAAAC 3'
[1065] SNP TSC0039960 was amplified using the following primer
set:
TABLE-US-00217 First primer: (SEQ ID NO: 611) 5'
TGTGGGTATTCTGAATTCCACAAAATGGACTAACACGC 3' Second primer: (SEQ ID
NO: 612) 5' ACGTCGTTCAGTGGGACATTAAAAGGCTC 3'
[1066] SNP TSC0418134 was amplified using the following primer
set:
TABLE-US-00218 First primer: (SEQ ID NO: 613) 5'
GGTTATGTGTCAGAATTCTGAAACTAGTTTGGAAGTAC 3' Second primer: (SEQ ID
NO: 614) 5' GCCTCAGTTTCGGGGACAGTTCTGAGGAC 3'
[1067] SNP TSC0603688 was amplified using the following primer
set:
TABLE-US-00219 First primer: (SEQ ID NO: 615) 5'
TGTAACACGGCCGAATTCCTCATTTGTATGAAATAGGT 3' Second primer: (SEQ ID
NO: 616) 5' AATCTAACTTGAGGGACCGGCACACACAC 3'
[1068] SNP TSC0129188 was amplified using the following primer
set:
TABLE-US-00220 First primer: (SEQ ID NO: 617) 5'
AGTGTCCCCTTAGAATTCGCAGAGACACCACAGTGTGC 3' Second primer: (SEQ ID
NO: 618) 5' TTTGCTACAGTCGGGACCCTTGTGTGCTC 3'
[1069] SNP TSC1103570 was amplified using the following primer
set:
TABLE-US-00221 First primer: (SEQ ID NO: 619) 5'
AGCACATCACTAGAATTCAATACCATGTGTGAGCTCAA 3' Second primer: (SEQ ID
NO: 620) 5' AATCCTGCTTCCGGGACCTAACTTTGAAC 3'
[1070] SNP TSC0813449 was amplified using the following primer
set:
TABLE-US-00222 First primer: (SEQ ID NO: 621) 5'
TTTCATTTTCTGGAATTCCTCTAATGATTTTCTGGAGC 3' Second primer: (SEQ ID
NO: 622) 5' CGTCGCCGCGTAGGGACTTTTTCTTCCAC 3'
[1071] SNP TSC0701940 was amplified using the following primer
set:
TABLE-US-00223 First primer: (SEQ ID NO: 623) 5'
TTACTTAATCCTGAATTCGAGAAAAGCCATGTTGATAA 3' Second primer: (SEQ ID
NO: 624) 5' TCATGGGTCGCTGGGACTTTGCCCTCTGC 3'
[1072] SNP TSC0087962 was amplified using the following primer
set:
TABLE-US-00224 First primer: (SEQ ID NO: 625) 5'
ACTAACAGCACTGAATTCATTTTACTATAATCTGCTAC 3' Second primer: (SEQ ID
NO: 626) 5' GTTAGCCGAGAAGGGACTGTCTGTGAAGC 3'
[1073] SNP TSC0660274 was amplified using the following primer
set:
TABLE-US-00225 First primer: (SEQ ID NO: 627) 5'
AAATATGCAGCGGAATTCGTAAGTGACCTATTAATAAC 3' Second primer: (SEQ ID
NO: 628) 5' GCGATGGTTACGGGGACAGCCAGGCAACC 3'
[1074] Each first primer had a biotin tag at the 5' end and
contained a restriction enzyme recognition site for EcoRI, and was
designed to anneal at a specified distance from the locus of
interest. This allows a single reaction to be performed for the
loci of interest, as each loci of interest will migrate at a
distinct position (based on annealing position of first primer).
The second primer contained a restriction enzyme recognition site
for BsmF I.
[1075] All loci of interest were amplified from the multiplexed
template DNA using the polymerase chain reaction (PCR, U.S. Pat.
Nos. 4,683,195 and 4,683,202, incorporated herein by reference). In
this example, the loci of interest were amplified in separate
reaction tubes but they could also be amplified together in a
single PCR reaction. For increased specificity, a "hot-start" PCR
was used. PCR reactions were performed using the HotStarTaq Master
Mix Kit supplied by QIAGEN (catalog number 203443).
[1076] The amount of multiplexed template DNA and primer per
reaction can be optimized for each locus of interest. One
microliter of the multiplexed template DNA eluted from the MinElute
column was used in the PCR reaction for each locus of interest, and
5 .mu.M of each primer was used. The twenty-nine SNPs described
above also were amplified from the maternal DNA (15 ng of DNA was
used in the PCR reaction; primer concentrations were as stated
above). Forty cycles of PCR were performed. The following PCR
conditions were used: [1077] (1) 95.degree. C. for 15 minutes and
15 seconds; [1078] (2) 37.degree. C. for 30 seconds; [1079] (3)
95.degree. C. for 30 seconds; [1080] (4) 57.degree. C. for 30
seconds; [1081] (5) 95.degree. C. for 30 seconds; [1082] (6)
64.degree. C. for 30 seconds; [1083] (7) 95.degree. C. for 30
seconds; [1084] (8) Repeat steps 6 and 7 thirty nine (39) times;
[1085] (9) 72.degree. C. for 5 minutes.
[1086] In the first cycle of PCR, the annealing temperature was
about the melting temperature of the 3' annealing region of the
second primers, which was 37.degree. C. The annealing temperature
in the second cycle of PCR was about the melting temperature of the
3' region, which anneals to the template DNA, of the first primer,
which was 57.degree. C. The annealing temperature in the third
cycle of PCR was about the melting temperature of the entire
sequence of the second primer, which was 64.degree. C. The
annealing temperature for the remaining cycles was 64.degree. C.
Escalating the annealing temperature from TM1 to TM2 to TM3 in the
first three cycles of PCR greatly improves specificity. These
annealing temperatures are representative, and the skilled artisan
will understand the annealing temperatures for each cycle are
dependent on the specific primers used.
[1087] The temperatures and times for denaturing, annealing, and
extension, can be optimized by trying various settings and using
the parameters that yield the best results. In this example, the
first primer was designed to anneal at various distances from the
locus of interest. The skilled artisan understands that the
annealing location of the first primer can be 5-10, 11-15, 16-20,
21-25, 26-30, 31-35, 36-40, 41-45, 46-50, 51-55, 56-60, 61-65,
66-70, 71-75, 76-80, 81-85, 86-90, 91-95, 96-100, 101-105, 106-110,
111-115, 116-120, 121-125, 126-130, 131-140, 140-160, 160-180,
180-200, 200-220, 220-240, 240-260. 260-280. 280-300, 300-350,
350-400, 400-450, 450-500, or greater than 500 bases from the locus
of interest.
Purification of Fragment of Interest
[1088] The PCR products were separated from the genomic template
DNA. Each PCR product was placed into a well of a Streptawell,
transparent, High-Bind plate from Roche Diagnostics GmbH (catalog
number 1 645 692, as listed in Roche Molecular Biochemicals, 2001
Biochemicals Catalog). Alternatively, the PCR products can be
pooled into a single well because the first primer was designed to
allow the loci of interest to separate based on molecular weight.
The first primers contained a 5' biotin tag so the PCR products
bound to the Streptavidin coated wells while the genomic template
DNA did not. The streptavidin binding reaction was performed using
a Thermomixer (Eppendorf) at 1000 rpm for 20 min. at 37.degree. C.
Each well was aspirated to remove unbound material, and washed
three times with 1.times.PBS, with gentle mixing (Kandpal et al.,
Nucl. Acids Res. 18:1789-1795 (1990); Kaneoka et al., Biotechniques
10:30-34 (1991); Green et al., Nucl. Acids Res. 18:6163-6164
(1990)).
Restriction Enzyme Digestion of Isolated Fragments
[1089] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in the Streptawells following the instructions supplied
with the restriction enzyme. After digestion, the wells were washed
three times with PBS to remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[1090] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase.
[1091] As demonstrated in Example 6, the sequence of both alleles
of a SNP can be determined by filling in the overhang with one
labeled nucleotide in the presence of the other unlabeled
nucleotides. The following components were added to each fill in
reaction: 1 .mu.l of fluorescently labeled ddGTP, 0.5 .mu.l of
unlabeled ddNTPs (40 .mu.M), which contained all nucleotides except
guanine, 2 .mu.l of 10.times. sequenase buffer, 0.25 .mu.l of
Sequenase, and water as needed for a 20 .mu.l reaction. The fill in
reaction was performed at 40.degree. C. for 10 min.
Non-fluorescently labeled ddNTP was purchased from Fermentas Inc.
(Hanover, Md.). All other labeling reagents were obtained from
Amersham (Thermo Sequenase Dye Terminator Cycle Sequencing Core
Kit, US 79565).
[1092] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme. Digestion was performed for 1 hour
at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[1093] After release from the streptavidin matrix, the sample was
loaded into a lane of a 36 cm 5% acrylamide (urea) gel
(BioWhittaker Molecular Applications, Long Ranger Run Gel Packs,
catalog number 50691). The sample was electrophoresed into the gel
at 3000 volts for 3 min. The gel was run for 3 hours on a
sequencing apparatus (Hoefer SQ3 Sequencer). The gel was removed
from the apparatus and scanned on the Typhoon 9400 Variable Mode
Imager. The incorporated labeled nucleotide was detected by
fluorescence.
[1094] Below a schematic of the 5' overhang for SNP TSC0838335 is
depicted. The entire sequence is not reproduced, only a portion to
depict the overhang (where R indicates the variable site).
TABLE-US-00226 10/14 5' TAA 3' ATT R A C A Overhang position 1 2 3
4
[1095] The observed nucleotides for TSC0838335 are adenine and
guanine on the 5' sense strand (herein depicted as the top strand).
The nucleotide in position three of the overhang corresponded to
cytosine, which is complementary to guanine. Labeled ddGTP can be
used to determine the sequence of both allele in the presence of
unlabeled dATP, dCTP, and dTTP.
[1096] The restriction enzyme BsmF I was used to create the 5'
overhang, which typically cuts 10/14 from the recognition site. At
times, BsmF I will cut 11/15 from the recognition site and generate
the following overhang:
TABLE-US-00227 11/15 5' TA 3' AT T R A C Overhang position 0 1 2
3
[1097] Position 0 in the overhang is thymidine, which is
complementary to adenine. Position 0 complementary to the overhang
was filled in with unlabeled dATP, and thus after the fill-in
reaction, the exact same molecules were generated whether the
enzyme cut at 10/14 or 11/15 from the recognition site. The DNA
molecules generated after the fill-in reaction are depicted
below:
TABLE-US-00228 G allele 10/14 5' TAA G* 3' ATT C A C A Overhang
position 1 2 3 4 G allele 11/15 5' TA A G* 3' AT T C A C Overhang
position 0 1 2 3 A allele 10/14 5' TAA A T G* 3' ATT T A C A
Overhang position 1 2 3 4 A allele 11/15 5' TA A A T G* 3' AT T T A
C Overhang position 0 1 2 3
[1098] The maternal template DNA amplified for TSC0838335 displayed
a single band that migrated at the expected position of the higher
molecular weight band, which corresponded to the "A" allele (see
FIG. 20, lane 1). The maternal template DNA was homozygous for
adenine at SNP TSC0838335.
[1099] However, in lane 2, amplification of the multiplexed
template DNA for TSC0838335 isolated from the plasma of the same
individual displayed two bands; a lower molecular weight band,
which corresponded to the "G" allele, and the higher molecular
weight band, which corresponded to the "A" allele. The template DNA
isolated from the plasma of a pregnant female contains both
maternal template DNA and fetal template DNA.
[1100] As seen in FIG. 20, lane 1, the maternal template DNA was
homozygous for adenine at this SNP (compare lanes 1 and 2). The "G"
allele represented the fetal DNA. Signals from the maternal
template DNA and the fetal template DNA clearly have been
distinguished. The "G" allele becomes a beacon for the fetal DNA
and can be used to measure the amount of fetal DNA present in the
sample. Additionally, once the percentage of fetal DNA in the
maternal plasma for a given sample has been determined, any
deviation from this percentage indicates a chromosomal abnormality.
This method provides the first non-invasive method for the
detection of fetal chromosomal abnormalities.
[1101] As seen in FIG. 20, lane 3, analysis of the maternal DNA for
SNP TSC0418134 generated a single band that migrated at the
expected position of the higher molecular weight band, which
corresponded to the adenine allele. Likewise, analysis of the
multiplexed template DNA isolated from the maternal plasma gave a
single band, which migrated at the expected position of the adenine
allele (see FIG. 20, lane 4). Both the maternal DNA and the fetal
DNA are homozygous for adenine at TSC0418134.
[1102] Below, a schematic of the 5' overhang for TSC0129188 is
depicted, wherein R indicates the variable site;
TABLE-US-00229 10/14 5' TCAT 3' AGTA R A C T Overhang position 1 2
3 4
[1103] The nucleotide upstream of the variable site (R) does not
correspond to guanine on the sense strand. Thus, the 5' overhang
generated by the 11/15 cutting properties of BsmF I will be
filled-in identically to the 5' overhang generated by the 10/14
cut. Labeled ddGTP in the presence of unlabeled dATP, dTTP, and
dCTP was used for the fill-in reaction. The DNA molecules generated
after the fill-in reaction are depicted below:
TABLE-US-00230 A allele 10/14 5' TCAT A T G* 3' AGTA T A C T
Overhang position 1 2 3 4 G allele 10/14 5' TCAT G* 3' AGTA C A C T
Overhang position 1 2 3 4
[1104] Analysis of the maternal DNA for SNP TSC0129188 gave a
single band that corresponded to the DNA molecules filled in with
ddGTP at position 1 complementary to the overhang, which
represented the "G" allele (see FIG. 20, lane 5). No band was
detected for adenine allele, indicating that the maternal DNA is
homozygous for guanine.
[1105] In contrast, analysis of the multiplexed template DNA from
the maternal plasma, which contains both maternal DNA, and fetal
DNA, gave two distinct bands (see FIG. 20, lane 6). The lower
molecular weight band corresponded to the "G" allele, while the
higher molecular weight corresponded to the "A" allele. The "A"
allele represents the fetal DNA. Thus, a method has been developed
that allows separation of maternal DNA and fetal DNA signals
without the added complexity of having to isolate fetal cells. In
addition, a sample of paternal DNA is not required to detect
differences between the maternal DNA and the fetal DNA.
[1106] Analysis of the maternal DNA for SNP TSC0501389 gave a
single band that migrated at the higher molecular weight position,
which corresponded to the "A" allele. No band was detected that
corresponded to the "G" allele. Similarly, analysis of the
multiplexed template DNA from the maternal plasma for SNP
TSC0501389 gave a single band that migrated at the higher molecular
weight position, which corresponded to the "A" allele. Both the
maternal template DNA and the fetal template DNA were homozygous
for adenine at SNP TSC0501389.
[1107] The maternal DNA and the template DNA from the plasma
originated from the same sample. One sample, which was obtained
through a non-invasive procedure, provided a genetic fingerprint
for both the mother and the fetus.
[1108] Of the twenty-nine SNPs for which the maternal template DNA
was homozygous, the fetal template DNA was heterozygous at two of
the twenty-nine SNPs. The fetal DNA was homozygous for the same
allele as the maternal template DNA at the remaining 27 SNPs (data
not shown). Comparing the homozygous allele of the maternal
template DNA and the plasma template DNA at a given SNP provides an
added level of quality control. It is not possible that the
maternal template DNA and the plasma template DNA are homozygous
for different alleles at the same SNP. If this is seen, it would
indicate that an error in processing had occurred.
[1109] The methods described herein demonstrate that the maternal
genetic signal can be separated and distinguished from the fetal
genetic signal in a maternal plasma sample. The above-example
analyzed SNPs located on chromosome 13, however any chromosome can
be analyzed including human chromosome 1, 2, 3, 4, 5, 6, 7, 8, 9,
10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X and Y and
fetal chromosomes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, X and Y.
[1110] In addition, the methods described herein can be used to
detect fetal DNA in any biological sample including but not limited
to cell, tissue, blood, serum, plasma, saliva, urine, tears,
vaginal secretions, umbilical cord blood, chorionic villi, amniotic
fluid, embryonic tissues, lymph fluid, cerebrospinal fluid, mucosa
secretions, peritoneal fluid, ascitic fluid, fecal matter, or body
exudates.
[1111] The methods described herein demonstrate that the percentage
of fetal DNA in the maternal sample can be determined by analyzing
SNPs wherein the maternal DNA is homozygous, and the DNA isolated
from the plasma of the pregnant female is heterozygous. The
percentage of fetal DNA can be used to determine if the fetal
genotype has any chromosomal disorders.
[1112] For example, if the percentage of fetal DNA present in the
sample is calculated to be 30% by analysis of chromosome 1
(chromosomal abnormalities involving chromosome 1 terminate early
in the pregnancy), then any deviation from 30% fetal DNA is
indicative of a chromosomal abnormality. For example, if upon
analysis of a SNP or multiple SNPs on chromosome 18, the percentage
of fetal DNA is higher than 30%, this would indicate that an
additional copy of chromosome 18 is present. The calculated
percentage of fetal DNA from any chromosome can be compared to any
other chromosome. In particular, the percentage of fetal DNA on
chromosome 13 can be compared to the percentage of fetal DNA on
chromosomes 18 and 21.
[1113] This analysis is assisted by knowledge of the expected ratio
of one allele to the other allele at each SNP. As discussed in
Example 9, not all heterozygous SNPs display ratios of 50:50.
Knowledge of the expected ratio of one allele to the other reduces
the overall number of variable sites that must be analyzed.
However, even without knowledge of the expected ratios for the
various SNPs, the percentage of fetal DNA can be calculated by
analyzing a large number of SNPs. When the sampling size of SNPs is
large enough, the statistical variation arising from the values of
the expected ratios will be eliminated.
[1114] In addition, heterozygous maternal SNPs also provide
valuable information. The analysis is not limited to homozygous
maternal SNPs. For example, if at a heterozygous SNP on maternal
DNA, the ratio of allele 1 to allele 2 is 1:1, then in the plasma
template DNA the ratio should remain 1:1 unless the fetal DNA
carries a chromosomal abnormality.
[1115] The above methods can also be used to detect mutations in
the fetal DNA including but not limited to point mutations,
transitions, transversions, translocations, insertions, deletions,
and duplications. As seen in FIG. 20, fetal DNA can readily be
distinguished from maternal DNA. The above methods can be used to
determine the sequence of any locus of interest for any gene.
Example 14
[1116] Plasma isolated from blood of a pregnant female contains
both maternal template DNA and fetal template DNA. As discussed
above, fetal chromosomal abnormalities can be determined by
analyzing SNPs wherein the maternal template DNA is homozygous and
the template DNA obtained from the plasma displays a heterozygous
pattern.
[1117] For example, assume SNP X can either be adenine or guanine,
and the maternal DNA for SNP X is homozygous for guanine. The
labeling method described in Example 6 can be used to determine the
sequence of the DNA in the plasma sample. If the plasma sample
contains fetal DNA, which is heterozygous at SNP X, the following
DNA molecules are expected after digestion with the type IIS
restriction enzyme BsmF I, and the fill-in reaction with labeled
ddGTP, unlabeled dATP, dTTP, and dCTP.
TABLE-US-00231 Maternal Allele 1 5' GGGT G* 3'CCCA C T C A Maternal
Allele 2 5' GGGT G* 3'CCCA C T C A Fetal Allele 1 5' GGGT G* 3'CCCA
C T C A Fetal Allele 2 5' GGGT A A G* 3'CCCA T T C A
[1118] Two signals are seen; one signal corresponds to the DNA
molecules filled in with ddGTP at position one complementary to the
overhang and the second signal corresponds to the DNA molecules
filled in with ddGTP at position three complementary to the
overhang. However, the maternal DNA is homozygous for guanine,
which corresponds to the DNA molecules filled in at position one
complementary to the overhang. The signal from the DNA molecules
filled in with ddGTP at position three complementary to the
overhang corresponds to the adenine allele, which represents the
fetal DNA. This signal becomes a beacon for the fetal DNA, and can
used to measure the amount of fetal DNA present in the plasma
sample.
[1119] There is no difference in the amount of fetal DNA from one
chromosome to another. For instance, the percentage of fetal DNA in
any given individual from chromosome 1 is the same as the
percentage of fetal DNA from chromosome 2, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X and Y. Thus, the
allele ratio calculated for SNPs on one chromosome can be compared
to the allele ratio for the SNPs on another chromosome.
[1120] For example, the allele ratio for the SNPs on chromosome 1
should be equal to the allele ratio for the SNPs on chromosomes 2,
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
21, 22, X, and Y. However, if the fetus has a chromosomal
abnormality, including but not limited to a trisomy or monosomy,
the ratio for the chromosome that is present in an abnormal copy
number will differ from the ratio for the other chromosomes.
[1121] To recapitulate the in vivo scenario of blood from a
pregnant female, maternal DNA was mixed with DNA isolated from her
child, who previously was diagnosed with Trisomy 21, in various
ratios to represent varying percentages of fetal DNA. For example,
to replicate the in vivo scenario of 50% fetal DNA in maternal
blood, equal amounts of maternal DNA were mixed with DNA isolated
from her child with Down's syndrome. The maternal DNA was analyzed
to identify homozygous SNPs, and these SNPs then were analyzed
using the mixture of 50% maternal DNA and 50% Down's syndrome DNA.
The ratio of allele 1 to allele 2 at heterozygous SNPs on
chromosome 13 was compared to the ratio of allele 1 to allele 2 at
heterozygous SNPs on chromosome 21.
[1122] Four different samples were analyzed: a sample with 100% of
the DNA from a child with Down syndrome; a sample with 75% DNA from
the child with Down syndrome and 25% DNA from the child's mother; a
sample with 50% DNA from the child with Down syndrome and 50% DNA
from the child's mother; and a sample with 40% DNA from the child
with Down syndrome and 60% DNA from the child's mother. The
maternal DNA was analyzed to identify homozygous SNPs. The DNA
isolated from the child with Down syndrome was genotyped to
identify heterozygous SNPs. Then, the samples were genotyped at
SNPs where the maternal DNA was homozygous and the DNA from the
child was heterozygous. For each sample, these SNPs were analyzed
ten times.
Collection of Blood Samples
[1123] An Internal Review Board approved study was designed to
allow collection of blood samples from children afflicted with
Down's syndrome and their parents. For this study, blood was
collected from the mother, the father, and the child with Down's
syndrome. Informed consent to collect blood from the child with
Down's syndrome was granted by the parents as well as the child.
Blood was collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284). The tubes were stored at 4.degree. C. until ready for
processing.
Isolation of Plasma and Maternal Cells
[1124] The blood was stored at 4.degree. C. until processing. The
tubes were spun at 1000 rpm for ten minutes in a centrifuge with
braking power set to zero. The tubes were spun a second time at
1000 rpm for ten minutes. The supernatant (the plasma) of each
sample was transferred to a new tube and spun at 3000 rpm for ten
minutes with the brake set to zero. The supernatant was transferred
to a new tube and stored at -80.degree. C. Approximately two
milliliters of the "buffy coat," which contains maternal cells, was
placed into a separate tube and stored at -80.degree. C.
Isolation of DNA
[1125] DNA was isolated from the plasma sample using the Qiagen
Midi Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA was eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also was used to isolate DNA from the maternal
cells contained in the "buffy coat." Maternal DNA and the plasma
DNA were isolated from the same tube of blood.
Identification of Maternal Homozygous SNPs
[1126] The maternal DNA was genotyped to identify homozygous SNPs.
Seven hundred and sixty-eight SNPs on chromosome 13 and 768 SNPs on
chromosome 21 were genotyped using the methods described in Example
6. Any number of SNPs can be analyzed, and the SNPs can be located
on human chromosome 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, X and Y. Preferably, the SNPs that
are genotyped have allele frequencies of 50:50, 60:40, 70:30,
80:20, or 90:10. As described in Example 8, the allele frequency of
any given SNP can be determined.
[1127] Details regarding the SNPs located on chromosome 13 and 21
can be found at the SNP consortium database, which can be accessed
via the internet at http://www.snp.cshl.org. The primers were
designed following the procedures set fourth in the Examples
described above, for example, in Examples 1, 2, 3, 5, and 6.
[1128] The first primers were designed so that after digestion with
a Type IIs enzyme, the products had different molecular weights as
described in Example 6. This allowed the amplified products to be
pooled, and run in a single lane of a gel.
[1129] For example, the first primer can be designed such that
after digestion a 30 base pair product is generated. Likewise, the
first primer of a different locus of interest can be designed such
that after digestion a 40 base pair product is generated. The first
primers can be designed so that in a single reaction, numerous loci
can be analyzed in one lane of a gel (30, 40, 50, 60, 70, 80, 90,
100, 110, and 120 base pair products can be run in a single lane).
The first primer can be designed to anneal any distance from the
locus of interest including but not limited to between 5-10, 10-25,
25-50, 50-75, 75-100, 100-150, 150-200, 200-250, 250-300, 300-350,
350-400, 400-450, 450-500, 500-550, 550-600, 600-650, 650-700,
700-750, 750-800, 800-850, 850-900, 900-950, 950-1000 and greater
than 1000 bases.
[1130] Amplification of the Loci of Interest
[1131] For each SNP that was genotyped, a PCR reaction was used to
amplify the loci of interest. The PCR reactions were performed in
96-well plates. The first and second primer (3 .mu.l of 1.25 .mu.M
stock concentration) for each SNP was distributed into a well of a
microtiter plate. Eight 96-well PCR plates were set-up for
chromosome 21 and eight 96-well plates were set-up for chromosome
13. After the primers had been distributed into the wells of the
microtiter plates, a mixture containing the genomic DNA and HotStar
PCR reagents was added to each well. Each PCR reaction contained 3
.mu.l of each primer, 7.5 .mu.l of HotStar Tag Master mix, 0.5
.mu.l of water, and 1 .mu.l of genomic DNA (10 ng/.mu.l).
[1132] The PCR cycling conditions were as follows:
[1133] (1) 95.degree. C. for 15 minutes and 15 seconds;
[1134] (2) 37.degree. C. for 30 seconds;
[1135] (3) 95.degree. C. for 30 seconds;
[1136] (4) 52.degree. C. for 30 seconds;
[1137] (5) 95.degree. C. for 30 seconds;
[1138] (6) 58.degree. C. for 30 seconds;
[1139] (7) 95.degree. C. for 30 seconds;
[1140] (8) Repeat steps 6 and 7 thirty seven (37) times;
[1141] (9) 72.degree. C. for 5 minutes.
Purification of Fragment of Interest
[1142] After the PCR reaction, 3 .mu.l of a PCR product generated
with a first primer designed to produce a 30 base pair product, 3
.mu.l of a PCR product generated with a first primer designed to
produce a 40 base pair product, 3 .mu.l of a PCR product generated
with a first primer designed to produce a 50 base pair product, 3
.mu.l of a PCR product generated with a first primer designed to
produce a 60 base pair product, 3 .mu.l of a PCR product generated
with a first primer designed to produce a 70 base pair product, 3
.mu.l of a PCR product generated with a first primer designed to
produce a 80 base pair product, 3 .mu.l of a PCR product generated
with a first primer designed to produce a 90 base pair product, 3
.mu.l of a PCR product generated with a first primer designed to
produce a 100 base pair product were mixed together in a well of a
Streptawell, transparent, High-Bind plate from Roche Diagnostics
GmbH (catalog number 1 645 692, as listed in Roche Molecular
Biochemicals, 2001 Biochemicals Catalog). The first primers
contained a 5' biotin tag so the PCR products bound to the
Streptavidin coated wells while the genomic template DNA did not.
The streptavidin binding reaction was performed using a Thermomixer
(Eppendorf) at 1000 rpm for 20 min. at 37.degree. C. Each well was
aspirated to remove unbound material, and washed three times with
1.times.PBS, with gentle mixing (Kandpal et al., Nucl. Acids Res.
18:1789-1795 (1990); Kaneoka et al., Biotechniques 10:30-34 (1991);
Green et al., Nucl. Acids Res. 18:6163-6164 (1990)).
Restriction Enzyme Digestion of Isolated Fragments
[1143] The purified PCR products were digested with the restriction
enzyme BsmF I, which binds to the recognition site incorporated
into the PCR products from the second primer. The digests were
performed in the Streptawells following the instructions supplied
with the restriction enzyme. After digestion, the wells were washed
three times with PBS to remove the cleaved fragments.
Incorporation of Labeled Nucleotide
[1144] The restriction enzyme digest with BsmF I yielded a DNA
fragment with a 5' overhang, which contained the SNP site or locus
of interest and a 3' recessed end. The 5' overhang functioned as a
template allowing incorporation of a nucleotide or nucleotides in
the presence of a DNA polymerase. As discussed in detail in Example
6, a single nucleotide labeled with one chemical moiety can be used
to determine the sequence at a SNP.
[1145] The amplified loci of interest were pooled into the
streptavidin-well based on size, and on the nucleotide used in the
fill-in reaction. The sequence of SNPs that were determined by
using a guanine nucleotide were pooled together. Likewise, the
sequence of SNPs that were determined by using an adenine
nucleotide were pooled together; the sequence of SNPs that were
determined by using a thymidine nucleotide were pooled together;
and the sequence of SNPs that were determined by using a cytosine
nucleotide were pooled together.
[1146] Thus, a typical fill-in reaction contained 8 amplified loci,
ranging in size of 30-120 base pair products; the sequence of all
eight was determined using a single nucleotide labeled with one
chemical moiety. Any number of amplified loci can be pooled
together.
[1147] The following components were added to each fill in
reaction: 1 .mu.l of fluorescently labeled dideoxynucleotide (ddGTP
for G fill-in reactions; ddATP for A fill-in reactions; ddTTP for
thymidine fill-in reactions; and ddCTP for cytosine fill-in
reactions), 0.5 .mu.l of unlabeled dNTPs (40 .mu.M), which
contained all nucleotides except the labeled nucleotide, 2 .mu.l of
10.times. sequenase buffer, 0.25 .mu.l of Sequenase, and water as
needed for a 20 .mu.l reaction.
[1148] The fill in reaction was performed at 40.degree. C. for 10
min. Non-fluorescently labeled dNTP was purchased from Fermentas
Inc. (Hanover, Md.). All other labeling reagents were obtained from
Amersham (Thermo Sequenase Dye Terminator Cycle Sequencing Core
Kit, US 79565).
[1149] After labeling, each Streptawell was rinsed with 1.times.PBS
(100 .mu.l) three times. The "filled in" DNA fragments were then
released from the Streptawells by digestion with the restriction
enzyme EcoRI, according to the manufacturer's instructions that
were supplied with the enzyme. Digestion was performed for 1 hour
at 37.degree. C. with shaking at 120 rpm.
Detection of the Locus of Interest
[1150] After release from the streptavidin matrix, the sample was
loaded into a lane of a 36 cm 5% acrylamide (urea) gel
(BioWhittaker Molecular Applications, Long Ranger Run Gel Packs,
catalog number 50691). The sample was electrophoresed into the gel
at 3000 volts for 3 min. The gel was run for 3 hours on a
sequencing apparatus (Hoefer SQ3 Sequencer). The gel was removed
from the apparatus and scanned on the Typhoon 9400 Variable Mode
Imager. The incorporated labeled nucleotide was detected by
fluorescence. The homozygous SNPs were identified.
Identification of Heterozygous SNPs with the Trisomy 21
Template
[1151] The DNA isolated from the individual with Down syndrome (the
child of the mother who was genotyped above) was analyzed to
identify heterozygous SNPs. The same seven hundred and sixty-eight
SNPs on chromosome 13 and the same 768 SNPs on chromosome 21 that
were analyzed with the maternal DNA were genotyped using the
methods described in Example 6. Any number of SNPs can be analyzed,
and the SNPs can be located on human chromosome 1, 2, 3, 4, 5, 6,
7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, X or
Y. Preferably, the SNPs that are genotyped have allele frequencies
of 50:50, 60:40, 70:30, 80:20, or 90:10. As described in Example 8,
the allele frequency of any given SNP can be determined.
[1152] The process for genotyping the SNPS with the DNA isolated
from the individual with Down syndrome was as described for the
maternal DNA. The heterozygous SNPs were identified.
[1153] SNPs that were homozygous for the maternal DNA and
heterozygous for the DNA isolated from the individual with Down
syndrome were further analyzed using samples that contained
mixtures of maternal DNA and Down syndrome DNA.
Generation of Samples Containing Maternal DNA and Down Syndrome
DNA
[1154] The DNA and the DNA obtained from her child, who has Down's
syndrome, were quantitated using a spectrophotometer. The maternal
DNA and the child's DNA were mixed together at various percentages
to represent the situation of circulating fetal DNA in the maternal
blood. The following percentages were analyzed: 100% Down's
syndrome DNA, 75% Down's syndrome DNA, 50% Down's syndrome DNA, and
40% Down's syndrome DNA.
[1155] The ratio at each heterozygous SNP was calculated by
dividing the value obtained for allele 1 by the value obtained for
allele 2. For example, if SNP X can either be adenine (A) or
guanine (G), the ratio at SNP X was calculated by dividing the
value obtained for adenine by the value obtained for guanine.
[1156] For the sample containing 100% Down syndrome DNA, sixty-two
SNPs on chromosome 13, which were homozygous with the maternal DNA
and heterozygous with the DNA isolated from the individual with
Down syndrome, were analyzed. For chromosome 21, forty-nine SNPs
were analyzed that were homozygous with the maternal DNA and
heterozygous with the DNA isolated from the individual with Down
syndrome.
[1157] The 62 SNPs on chromosome 13 and 49 SNPs on chromosome 21
were analyzed ten separate times. As shown in Table XX, for each of
the ten trials, the ratio of allele 1 to allele 2 on chromosome 13
was approximately 1.0 as expected. For chromosome 13, there is one
copy of allele 1 and one copy of allele 2. The average of the ten
trials was 1.051 with a standard deviation of 0.085.
[1158] With a Trisomy 21, there are two copies of one allele, which
are usually inherited from the mother, and one copy of the other
allele. The expected ratio is approximately 0.5 (one copy of allele
1/two copies of allele 2). As shown in Table XX, the ratio for
chromosome 21 varied from a low of 0.462 to a high of 0.634. For
every trial, the ratio obtained for chromosome 21 was significantly
distinct from the ratio obtained at chromosome 13. The average
ratio for the ten trials was 0.531 with a standard deviation of
0.049.
[1159] The experiment was repeated ten times so that a true
statistical measurement could be obtained. If ten different genetic
samples were used, the SNPs that fit the criteria (maternal
homozygous, Down syndrome child heterozygous) would be different,
making it difficult to compare from sample to sample.
[1160] Statistical analysis revealed a confidence value of 99.9%
that the ratios obtained on chromosome 13 and on chromosome 21
represented true differences, rather than random numerical
fluctuations in value. The Ravgen method identified the presence of
the chromosomal abnormality.
[1161] For the sample containing 75% Down syndrome DNA and 25%
maternal DNA, sixty two SNPS on chromosome 13 and fifty SNPs on
chromosome 21 were analyzed, unless stated otherwise. For various
trials, not all the SNPS could be quantitated because the bands
corresponding to certain SNPs were faint. This may have been caused
by poor PCR amplification, poor binding to the streptavidin plate,
or a weak fill-in reaction.
[1162] For trial 3, 61 SNPs on chromosome 13 were analyzed. For
trail 4, 49 SNPs were analyzed on chromosome 21. With regard to
trial 5, 47 SNPs on chromosome 21 were analyzed and 61 SNPs on
chromosome 13. For trial 7, 49 SNPs were analyzed on chromosome 21
and 61 SNPs on chromosome 13. For trial 8, 49 chromosomes were
analyzed on chromosome 21, and 59 SNPs were analyzed on chromosome
13. For trials 9 and 10, 59 SNPs on chromosome 13 were
analyzed.
[1163] The expected ratio on chromosome 13 for a heterozygous SNP
is 0.6. If the maternal chromosomes both contain an adenine
nucleotide, and the Down syndrome genome is comprised of one
chromosome with an adenine nucleotide and one chromosome with a
guanine nucleotide, then the ratio of G:A is 0.75/(0.75 (Down
syndrome A allele)+0.25+0.25 (maternal A alleles)), which is 0.6.
For the ten trials, the ratios obtained for chromosome 13 varied
from 0.567 to 0.645. The average for the ten trials was 0.609 with
a standard deviation of 0.032 (see Table XX).
[1164] The expected ratio for chromosome 21 in a Trisomy condition
is 0.375. If the maternal chromosomes both contain an adenine
nucleotide, and the Down syndrome genome is comprised of two
chromosomes with an adenine nucleotide and one chromosome with a
guanine nucleotide, then the ratio of G:A is 0.75/(0.75+0.75 (Down
syndrome A alleles)+0.25+0.25 (maternal A alleles)), which is
0.375.
[1165] For the ten trials, the ratios obtained for chromosome 21
varied from 0.350 to 0.4125, with an average of 0.384 and a
standard deviation of 0.017 (see Table XX). Statistical analysis
revealed a confidence value of 99.9% that the ratios obtained on
chromosome 13 and on chromosome 21 represented true differences,
rather than random numerical fluctuations in value. The Ravgen
method identified the presence of the chromosomal abnormality in
the presence of 25% maternal DNA.
[1166] With regard to the sample containing 50% Down syndrome DNA,
46 SNPs on chromosome 13 and 35 SNPs on chromosome 21 were
analyzed, unless stated otherwise. For trial 1, 45 SNPs on
chromosome 13 were analyzed. For trail 2, 44 SNPs on chromosome 13
were analyzed. For trial 3, 42 SNPs on chromosome 13 were analyzed.
For trial 4, 44 SNPs on chromosome 13 and 34 SNPs on chromosome 21
were analyzed. For trial 5, 34 SNPs on chromosome 21 were analyzed.
For trials 7 and 8, 44 and 41 SNPs on chromosome 13, respectively,
were analyzed. For trial 9, 44 SNPs on chromosome 13 and 34 SNPs on
chromosome 21 were analyzed. For trial 10, 44 SNPs on chromosome 13
were analyzed.
[1167] The expected ratio at a heterozygous SNP on chromosome 13
for the 50% sample is 0.33. If the maternal chromosomes both
contain an adenine nucleotide, and the Down syndrome genome is
comprised of one chromosome with an adenine nucleotide and one
chromosome with a guanine nucleotide, then the ratio of G:A is
0.50/(0.50 (Down syndrome A allele)+0.50+0.50 (maternal A
alleles)), which is 0.33. For the ten trials, the ratios obtained
for chromosome 13 varied from 0.302 to 0.347. The average for the
ten trials was 0.324 with a standard deviation of 0.0.13 (see Table
XX).
[1168] The expected ratio for chromosome 21 in a Trisomy condition
is 0.25. If the maternal chromosomes both contain an adenine
nucleotide, and the Down syndrome genome is comprised of two
chromosomes with an adenine nucleotide and one chromosome with a
guanine nucleotide, then the ratio of G:A is 0.50/(0.50+0.50 (Down
syndrome A alleles)+0.50+0.50 (maternal A alleles)), which is
0.25.
[1169] For the ten trials, the ratios obtained for chromosome 21
varied from 0.230 to 0.275, with an average of 0.244 and a standard
deviation of 0.015 (see Table XX). Statistical analysis revealed a
confidence value of 99.1% that the ratios obtained on chromosome 13
and on chromosome 21 represented true differences, rather than
random numerical fluctuations in value. The Ravgen method
identified the presence of the chromosomal abnormality in the
presence of 50% maternal DNA.
[1170] For the sample containing 40% Down syndrome DNA, 60 SNPs on
chromosome 13 and 48 SNPs on chromosome 21 were analyzed, unless
stated otherwise. For trial 1, 47 SNPs on chromosome 21 were
analyzed. For trials 2-4, 59 SNPs on chromosome 13 and 47 SNPs on
chromosome 21 were analyzed. For trials 5 and 6, 46 SNPs on
chromosome 21 were analyzed. For trail 7, 58 SNPs on chromosome 13
were analyzed. For trial 8, 46 SNPs on chromosome 21 were analyzed
and for trials 9 and 10, 47 SNPs on chromosome 21 were
analyzed.
[1171] The expected ratio at a heterozygous SNP on chromosome 13
for the 40% Down syndrome DNA sample is 0.25. If the maternal
chromosomes both contain an adenine nucleotide, and the Down
syndrome genome is comprised of one chromosome with an adenine
nucleotide and one chromosome with a guanine nucleotide, then the
ratio of G:A is 0.40/(0.40 (Down syndrome A allele)+0.60+0.60
(maternal A alleles)), which is 0.25. For the ten trials, the
ratios obtained for chromosome 13 varied from 0.254 to 0.285. The
average for the ten trials was 0.269 with a standard deviation of
0.009 (See Table XX).
[1172] The expected ratio for chromosome 21 in a Trisomy condition
is 0.20. If the maternal chromosomes both contain an adenine
nucleotide, and the Down syndrome genome is comprised of two
chromosomes with an adenine nucleotide and one chromosome with a
guanine nucleotide, then the ratio of G:A is 0.40/(0.40+0.40 (Down
syndrome A alleles)+0.60+0.60 (maternal A alleles)), which is
0.20.
[1173] For the ten trials, the ratios obtained for chromosome 21
varied from 0.216 to 0.249, with an average of 0.23 and a standard
deviation of 0.011 (see Table XX). Statistical analysis revealed a
confidence value of 94.3% that the ratios obtained on chromosome 13
and on chromosome 21 represented true differences, rather than
random numerical fluctuations in value. The Ravgen method
identified the presence of the chromosomal abnormality in the
presence of 60% maternal DNA.
[1174] The presence of the Trisomy 21 condition was identified with
the Ravgen method in numerous samples that contained various
percentages of abnormal DNA. Each percentage of abnormal DNA was
analyzed ten separate times and each time, the presence of the
abnormal condition was identified. The ratio of allele 1 to allele
2 at multiple heterozygous SNPs on chromosome 13 was calculated,
and the ratios were averaged. The same was done with the SNPs
located on chromosome 21. The ratio obtained for the heterozygous
SNPs on chromosome 13 was statistically different from the ratio
obtained on chromosome 21. The ratios obtained on both chromosome
13 and 21 were near the mathematically predicted values.
[1175] In this example, the confidence interval for the samples
with 100% Down syndrome DNA and 75% Down syndrome DNA was 99.9%,
and the confidence interval for the sample with 50% Down syndrome
DNA was 99.1%, which is about the accuracy reported for
amniocentesis. The confidence interval for the sample containing
40% Down syndrome DNA was 94.3%, which is more accurate than
currently marketed non-invasive tests for prenatal diagnostics.
[1176] As discussed above, about 60 SNPs on chromosome 13 and 50
SNPs on chromosome 21 were analyzed. To increase the confidence
interval for samples containing 40% fetal DNA or lower, a larger
number of SNPs can be analyzed. The Ravgen method provides a highly
accurate, cost-effective way to sequence DNA, so sequencing a
larger number of SNPs is not difficult. The accuracy of the test is
determined by the number of SNPs that are sequenced. For higher
accuracy with samples that contain lower percentages of DNA, more
SNPs can be analyzed. Alternatively, the methods described in this
application can be used to ensure that the samples contain a higher
percentage of fetal DNA.
[1177] In this example, a sample containing 40% Down syndrome DNA,
which represented the fetal DNA in the maternal blood, was
analyzed. Maternal blood samples with any percentage of fetal DNA
can be analyzed including but not limited to 0.0001-1%, 1-10%,
10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%, 70-80%, 80-90%, and
90-100%.
TABLE-US-00232 TABLE XX The Ravgen method identifies chromosomal
abnormalities in Samples containing 40% Down syndrome DNA Expected
Expected Ratio at Chrom. 13 21 21 13 21 21 100% DS DNA ~75% DS DNA
Trial 1 0.959 0.5708 0.49 0.637 .3764 0.389 Trial 2 0.916 0.5024
0.48 0.567 .3894 0.362 Trial 3 1029 0.4616 0.51 0.651 .3707 0.394
Trial 4 0.967 0.5123 0.491 0.580 .3901 0.367 Trial 5 1.037 0.6339
0.51 0.645 .4125 0.392 Trial 6 1.111 0.5425 0.53 0.645 .3743 0.392
Trial 7 1.154 0.495 0.54 0.594 .3974 0.373 Trial 8 1.135 0.5276
0.532 0.583 .3901 0.368 Trial 9 1.148 0.5619 0.534 0.579 .3899
0.367 Trial 10 1.057 0.4976 0.52 0.609 .350 0.378 AVG. 1.051 .531
0.512 .609 0.384 0.378 STDEV .085 .049 .032 .017 ~50% DS DNA ~40%
DS DNA Trial 1 0.347 0.275 0.258 0.277 0.239 0.217 Trial 2 0.3.16
0.237 0.24 0.265 0.249 0.21 Trial 3 0.338 0.247 0.253 0.266 0.227
0.21 Trial 4 0.331 0.264 0.249 0.254 0.216 0.202 Trial 5 0.330
0.241 0.248 0.274 0.246 0.215 Trial 6 0.324 0.240 0.244 0.268 0.22
0.211 Trial 7 0.318 0.233 0.241 0.275 0.227 0.216 Trial 8 0.302
0.230 0.231 0.258 0.228 0.21 Trial 9 0.315 0.238 0.240 0.285 0.231
0.222 Trial 10 0.318 0.235 0.241 0.266 0.218 0.21 AVG. 0.324 0.244
0.244 0.269 0.23 0.212 STDEV 0.013 0.015 0.009 0.011
Example 15
[1178] As discussed in Example 4 above, the use of cell lysis
inhibitors, cell membrane stabilizers, or cross-linking reagents
can be used to increase the percentage of fetal DNA in the maternal
blood. In this example, methods for the isolation of free fetal DNA
are disclosed, which minimize the amount of maternal cell lysis.
The effect of formalin on sixty-nine (69) maternal blood samples
from twenty-seven clinical practices located in sixteen different
states was analyzed. Formalin was added to all samples collected
from the pregnant women, and the percentage of fetal DNA was
calculated using serial dilution analysis followed by PCR. A
genetic marker on the Y chromosome was used to calculate the
percent of fetal DNA.
Collection of Blood Samples
[1179] In accordance with an IRS approved study, blood samples were
collected from pregnant women after informed consent had been
granted. Blood samples were received from 27 different clinical
sites operating in 16 different states located throughout the U.S.
Blood samples were collected from both women carrying male and
female fetuses, however, here, we report results obtained from
woman carrying male fetuses, as the Y chromosome is the accepted
marker when quantitating percentages of fetal DNA.
[1180] Blood is collected by any method or process that results in
a substantial increase in the ratio of fetal DNA/maternal DNA in
the resulting serum or plasma after appropriate processing. As used
herein, a substantial increase in the ratio of fetal DNA/maternal
DNA is that which can be detected by the methods as described
herein. Such methods or processes typically result in a substantial
increase in the ratio of fetal DNA/maternal DNA of about 5%, 10%,
15%, 20%, 30%, 50%, 70%, 80%, 100% or more of the ratio of fetal
DNA/maternal DNA found in blood samples collected by standard
procedures.
[1181] In other embodiments, blood is collected by any method or
process that results in a substantial increase in the amount of
free fetal DNA compared to the amount of total DNA recovered or
detected in the resulting serum or plasma after processing. Such
methods or processes typically result in a substantial increase so
the fetal DNA recovered or detected is about 10%, 15%, 20%, 25%,
30%, 40%, 50% or more of the total DNA recovered or detected in the
processed plasma or serum sample.
[1182] All clinical sites were provided with a kit used for the
venipuncture procedure, which included 21 gauge needles, 9 ml EDTA
Vacuette tubes (catalog number NC9897284) a syringe containing
0.225 ml of 10% neutral buffered solution containing formaldehyde
(4% w/v), an icepack, and a shipping container. The clinical sites
were instructed to add the formaldehyde immediately after drawing
the blood and to gently invert the tubes.
[1183] The methods or processes of collecting blood samples may
also include other steps that result in lessened or reduced cell
lysis. For instance, blood collection devices may be modified to
decrease cell lysis due to sheer forces in the collection needle,
syringe or tubes used. For instance, needles of large gauge may be
employed to reduce cell sheering or vacutainer tubes may be
modified to reduce the velocity of blood flow.
Isolation of Plasma
[1184] Any method may be used to isolate plasma from the cell
components of blood after collection but methods wherein cell lysis
is substantially prevented, reduced or inhibited are preferred. The
blood was stored at 4.degree. C. until processing. Methods for
isolation of the plasma were implemented to reduce the amount of
maternal cell lysis. The tubes were spun at 1000 rpm for ten
minutes in a centrifuge with braking power and acceleration power
set at zero to substantially prevent, reduce or inhibit cell lysis
and or mixing of blood cell components into the plasma. The tubes
were spun a second time at 1000 rpm for ten minutes with braking
power (centrifuge stopped by natural deceleration) and acceleration
power set to zero. The supernatant (the plasma) of each sample was
transferred carefully to a new tube and spun at 3000 rpm for ten
minutes with the brake and acceleration power set at zero. The
supernatant (the plasma) of each sample was collected via
procedures to substantially prevent mixing of cell components into
the plasma. Great care was taken to ensure that the buffy-coat was
not disturbed. A percentage of the supernatant can be left in the
tube including but not limited to 0.001-1%, 1-10%, 10-20%, 20-30%,
30-40%, 40-50%, 50-60%, 60-70%, 70-80% or greater than 80%. In this
example, about 0.5 ml of the supernatant was left in the tube to
ensure that the buffy-coat was not disturbed. The supernatant was
transferred to a new tube and stored at -80.degree. C.
Isolation of DNA
[1185] DNA was isolated from the plasma sample using the Qiagen
Midi Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183), DNA was eluted in 100 .mu.l of distilled water.
However, any method of DNA isolation can be used including cesium
chloride gradients, gradients, sucrose gradients, glucose
gradients, centrifugation protocols, boiling, Qiagen purification
systems, QIA DNA blood purification kit, HiSpeed Plasmid Maxi Kit,
QIAfilter plasmid kit, Promega DNA purification systems, MangeSil
Paramagnetic Particle based, systems, Wizard SV technology, Wizard
Genomic DNA purification kit, Amersham purification systems, GFX
Genomic Blood DNA purification kit, Invitrogen Life Technologies
Purification Systems, CONCERT purification system, Mo Bio
Laboratories purification systems, UltraClean BloodSpin Kits, and
UlraClean Blood DNA Kit. The skilled artisan understands that the
manufacturer's protocols can modified to increase the yield of DNA.
For example, the Qiagen Midi Kit for purification of DNA recommends
the use of 1.times.AL buffer. However, any concentration of AL
buffer may be used if the yield of DNA increases including but not
limited to 0.1-0.5.times.AL buffer, 0.5-1.times.AL buffer,
1.times.-2.times.AL buffer, 2-3.times.AL buffer, 3-4.times.AL
buffer, 4-5.times.AL buffer, and greater than 5.times.AL buffer.
The skilled artisan understands that the modifications and
manipulations of the reagents are not limited to AL buffer.
Quantification of Percentage of Fetal DNA
[1186] The percentage of fetal DNA present in the maternal plasma
sample was calculated using serial dilution analysis followed by
PCR. Two different sets of primers were used: one primer set was
specific for the Y chromosome, and thus specific for fetal DNA, and
the other primer set was designed to amplify the cystic fibrosis
gene, which is present on both maternal template DNA and fetal
template DNA.
Primer Design:
[1187] The following primers were designed to amplify the SRY gene
on the Y chromosome:
TABLE-US-00233 Upstream primer: 5' TGGCGATTAAGTCAAATTCGC 3' (SEQ ID
NO: 263) Downstream primer: 5' CCCCCTAGTACCCTGACAATGTATT 3' (SEQ ID
NO: 264)
[1188] The following primers were designed to amplify the cystic
fibrosis gene:
TABLE-US-00234 Upstream primer: 5' CTGTTCTGTGATATTATGTGTGGT 3' (SEQ
ID NO: 265) Downstream primer: 5' AATTGTTGGCATTCCAGCATTG 3' (SEQ ID
NO: 266)
PCR Reaction
[1189] The SRY gene and the cystic fibrosis gene were amplified
from the template genomic DNA using PCR (U.S. Pat. Nos. 4,683,195
and 4,683,202). For increased specificity, a "hot-start" PCR was
used. PCR reactions were performed using the HotStarTaq Master Mix
Kit supplied by Qiagen (Catalog No. 203443). For amplification of
the SRY gene, the DNA eluted from the Qiagen purification column
was diluted serially 1:2. For amplification of the cystic fibrosis
gene, the DNA eluted from the Qiagen purification column was
diluted 1:4, and then serially diluted 1:2. The following
components were used for each PCR reaction: 8 .mu.l of template DNA
(diluted or undiluted), 1 .mu.l of each primer (5 .mu.M), 10 .mu.l
of HotStar Taq mix. The following PCR conditions were used:
[1190] (1) 95.degree. C. for 15'
[1191] (2) 94.degree. C. for 1'
[1192] (3) 54.degree. C. for 15''
[1193] (4) 72.degree. C. for 30''
[1194] (5) Repeat steps 2-4 for 45 cycles.
[1195] (6) 10' at 72.degree. C.
[1196] Amplification of the SRY gene was performed using the
following templates: undiluted, diluted 1:2, diluted 1:4, diluted
1:8, diluted 1:16, diluted 1:32, diluted 1:64, diluted 1:128,
diluted 1:256, and diluted 1:512. Amplification of the cystic
fibrosis gene was performed using the following templates: diluted
1:4, diluted 1:8, diluted 1:16, diluted 1:32, diluted 1:64, diluted
1:128, diluted 1:256, diluted 1:512, diluted 1:1024, diluted
1:2048, and diluted 1:4096.
[1197] The percent of fetal DNA present in the maternal plasma was
calculated using the following formula:
% fetal DNA=(amount of SRY gene/amount of cystic fibrosis
gene)*2*100.
[1198] The amount of SRY gene was represented by the highest
dilution value in which the gene was amplified. Likewise, the
amount of cystic fibrosis gene was represented by the highest
dilution value in which it was amplified. The formula contains a
multiplication factor of two (2), which is used to normalize for
the fact that there is only one copy of the SRY gene (located on
the Y chromosome), while there are two copies of the cystic
fibrosis gene.
[1199] The effect of formalin on sixty-nine (69) maternal blood
samples collected from twenty-seven clinical practices located in
sixteen different states, spanning from Washington to Massachusetts
is shown in Table XXI. In this study, formalin was added to all
samples collected from the pregnant women, and the percentage of
fetal DNA was calculated using serial dilution analysis followed by
PCR. The serial dilutions and PCR amplifications were performed by
four different scientists over a period of five months. The samples
were collected from women at gestational ages ranging from 11 weeks
to 28 weeks, with the majority of women between 16-19 weeks of
gestation. A summary is provided in Table XXIII.
[1200] The average percentage of free fetal DNA for the 69 samples
analyzed in the maternal blood was 33.6%. Lo et al. reported fetal.
DNA concentrations of 3.4% in woman in late first to mid-second
trimester, which was the gestational age of the majority of women
in this study. Thus, the addition of formalin led to approximately
a ten-fold increase in the average percentage of fetal DNA.
[1201] While the calculated percentage of fetal DNA in maternal
blood is impressive, it is also informative to examine the range of
the percentages of fetal DNA observed in this study. About six
percent of the women (4/69) had 3.125% of free fetal DNA in the
maternal blood, which was the lowest percentage of fetal. DNA
observed in this study. Another 10.2% of women had 6.25% fetal DNA,
which represents a two-fold increase over the reported average in
the literature. The total number of women who had less than 10%
fetal DNA in the maternal blood was only 16.0%.
[1202] Fifty-eight percent of the women in this study had a
percentage of fetal DNA of 25% or greater. Importantly, 26.0% of
the women had fifty percent or greater fetal DNA in the maternal
blood. Fetal DNA percentages of this magnitude have not been
reported, and represent a new tool to the field of prenatal
genetics.
[1203] There were four samples collected from women at the
gestational age of eleven weeks. The percentages of fetal DNA in
the maternal blood samples were as follows: two samples at 12.5%;
one sample at 25%; and one sample at greater than 50%. Thus, the
effect of formalin on the percentages of fetal DNA was observed
with samples collected from women in early as well as later
gestational periods.
[1204] The effect of stabilizing cell membranes and reducing the
release of free DNA was not limited to formalin. We have tested
several different types of agents, and combinations of agents, that
prevent cell lysis and/or stabilize cell membranes, such as
glutaraldehyde, and have seen that these agents also reduce the
amount of free DNA in the blood sample (data not shown).
[1205] The above described methods may also include steps of adding
an agent to the blood sample at the time or near to the time of
collection to substantially inhibit or impede cell lysis or
stabilize cell membranes. Any number of agents that impede cell
lysis or stabilize cell membranes or cross-link cell membranes can
be added to the maternal blood samples including but not limited to
formaldehyde, and derivatives of formaldehyde, formalin,
glutaraldehyde, and derivatives of glutaraldehyde, crosslinkers,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP,
DTME, DTSSP, EGS, HBVS, sulfa-BSOCOES, Sulfa-DST, Sulfo-EGS or the
compounds listed in Table XXIII. Additional cross-linkers that can
be used are found at the following website:
www.piercenet.com/products/.
[1206] An agent that stabilizes cell membranes may be added to the
maternal blood sample to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, mexitil, dilantin, sodium hyaluronate, or polaxamer
188.
[1207] Any concentration of agent that stabilizes cell membranes,
impedes cell lysis or cross-link cell membranes can be added. In a
preferred embodiment, the agent that stabilizes cell membranes,
impedes cell lysis, or cross-links cell membranes is added at a
concentration that does not impede or hinder subsequent
reactions.
[1208] While impressive percentages of free fetal DNA in maternal
blood samples have been reported, it is thought that higher
percentages can be achieved by carefully explaining the importance
of the formalin to the physicians. Samples randomly were checked
for the presence of formalin and found that about ten percent of
the samples did not receive formalin. In addition, aggregates were
observed in another ten percent of the samples suggesting that the
formalin had not been thoroughly mixed with the collected blood.
Thus, while the addition of formalin produced an impressive effect,
it is likely that under controlled conditions, the percentage of
free fetal DNA may be higher.
[1209] In addition, we believe that procedures to minimize
hemolysis during the venipuncture procedure and temperature
controlled shipping containers (specimens were shipped in a
Styrofoam container with ice pack, but there was variation in
temperature because samples were shipped from varying distances)
may cause a further increase in the percentage of free fetal DNA.
Needles designed to reduce hemolysis can be used during the
venipuncture procedure.
[1210] Also, we hypothesized that procedures for carefully
isolating the plasma would help to ensure a minimal amount of
maternal DNA in the sample. We implemented procedures, as described
above, to reduce cell lysis, such as gentle centrifugation
parameters, and allowed the rotors to stop without external force
(no brake). Also, we carefully removed the supernatant containing
the plasma DNA from the buffy-coat, which contains maternal DNA.
These procedures coupled with the addition of formalin to prevent
cell lysis resulted in a tremendous increase in the percentage of
fetal DNA.
TABLE-US-00235 TABLE XXI Formalin increases the percentage of free
fetal DNA in blood samples collected at numerous clinical sites
from women at various stages of gestation. Wks Sample Gestation
Fetal Genomes/ml % Fetal DNA 1 16 80 25 2 19 1066 >50 3 17 52 50
4 22 166 25 5 32 457 50 6 19 400 100 7 18 800 100 8 17 100 50 9 16
50 25 10 17 25 12.5 11 16 94.74 12.5 12 16 34.60 50 13 16 22.5 25
14 17 50 12.5 15 17 26.48 12.5 16 17 45.00 25 17 17 94.7 100 18 17
28.13 6.25 19 19 28.13 25 20 20 11.25 12.5 21 15 11.25 12.5 22 11
16.66 12.5 23 18 13.23 25 24 18 12.50 6.25 25 16 112.50 100 26 17
124.13 25 27 14 90.00 50 28 11 100.00 100 29 18 232.00 100 30 19
626.00 100 31 19 112.50 100 32 16 423.50 100 33 16 423.50 25 34 11
105.88 25 35 16 49.60 3.1 36 11 11.84 12.5 37 16 120.00 25 38 18
342.90 100 39 17 51.43 25 40 18 225.00 6.25 41 17 400.00 12.5 42 28
180.00 25 43 17 20.45 12.5 44 18 25.73 25 45 16 68.68 3.1 46 17
218.18 25 47 15 75.00 6.25 48 16 40.58 3.1 49 17 100.00 25 50 17
14.06 12.5 51 22 22.50 12.5 52 15 28.13 12.5 53 17 50.00 3.125 54
18 58.00 50 55 14 100.00 25 56 16 58.08 25 57 16 13.64 12.5 58 16
25.00 6.25 59 20 45.00 25 60 16 23.69 12.5 61 18 5.92 6.25 62 15
28.13 6.25 63 17 50.00 25 63 64 16 360.00 50 65 16 25.00 12.5 66 16
48.65 25 67 16 47.38 12.5 68 14 26.45 50 69 17 124.15 25 Average 17
131.15 33.6
TABLE-US-00236 TABLE XXII Formalin increases the percentage of free
fetal DNA in blood samples collected at numerous clinical sites
from women at various stages of gestation. % Fetal DNA 3.125 6.25
12.5 25 50 Over 50% Number 4 7 18 22 7 11 Women (69) % 5.8 10.1
26.1 31.9 10.2 15.9
TABLE-US-00237 TABLE XXIII A representative list of cross-linkers
that can be used to impede maternal cell lysis. Cross-Linker
Abbreviation succinimidyl acetylthioacetate SATA succinimidyl
trans-4-(maleimidylmethyl) cyclohexane-1-carboxylate SMCC
succinimidyl 3-(2-pyridyldithio)propionate SPDP
N-((2-pyridyldithio)ethyl)-4-azidosalicylamide PEAS; AET
4-azido-2,3,5,6-tetrafluorobenzoic acid, succinimidyl ester ATFB,
SE 4-azido-2,3,5,6-tetrafluorobenzoic acid, STP ester, sodium salt
ATFB, STP ester 4-azido-2,3,5,6-tetrafluorobenzyl amine,
hydrochloride benzophenone-4-isothiocyanate
benzophenone-4-maleimide 4-benzoylbenzoic acid, succinimidyl ester
Disuccinimidylsuberate DSS Dithiobis(succinimidylpropionate) DSP
3,3'-Dithiobis(sulfosuccinimidylpropionate) DTSSP
Bis[2-(sulfosuccinimdooxycarbonyloxy)ethyl]sulfone SULFO BSOCOES
Bis[2-(succinimdooxycarbonyloxy)ethyl]sulfone BSOCOES
Disulfosuccinimdyltartrate SULFO DST Disuccinimdyltartrate DST
Ethylene glycolbis(succinimidylsuccinate) SULFO EGS Ethylene
glycolbis(sulfosuccinimidylsuccinate) EGS
1,2-Di[3'-(2'-pyridyldithio)propionamido]butane DPDPB
Bis(sulfosuccinimdyl)suberate BSSS
Succinimdyl-4-(p-maleimidophenyl)butyrate SMPB
Sulfosuccinimdyl-4-(p-maleimidophenyl)butyrate SULFO SMPB
3-Maleimidobenzoyl-N-hydroxysuccinimide ester MBS
3-Maleimidobenzoyl-N-hydroxysulfosuccinimide ester SULFO MBS
N-Succinimidyl(4-iodoacetyl)aminobenzoate SIAB
N-Sulfosuccinimidyl(4-iodoacetyl)aminobenzoate SULFO SIAB
Succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate SMCC
Sulfosuccinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate
SULFO SMCC
Succinimidyl-6-[3-(2-pyridyldithio)propionamido)hexanoate NHS LC
SPDP Sulfosuccinimidyl-6-[3-(2-pyridyldithio)propionamido)hexanoate
SULFO NHS LC SPDP N-Succinimdyl-3-(2-pyridyldithio)propionate SPDP
N-Hydroxysuccinimidylbromoacetate NHS BROMOACETATE
N-Hydroxysuccinimidyliodoacetate NHS IODOACETATE
4-(N-Maleimidophenyl)butyric acid hydrazide hydrochloride MPBH
4-(N-Maleimidomethyl)cyclohexane-1-carboxylic acid hydrazide MCCH
hydrochloride m-Maleimidobenzoic acid hydrazidehydrochloride MBH
N-(epsilon-Maleimidocaproyloxy)sulfosuccinimide SULFO EMCS
N-(epsilon-Maleimidocaproyloxy)succinimide EMCS
N-(p-Maleimidophenyl)isocyanate PMPI N-(kappa-Maleimidoundecanoic
acid) hydrazide KMUH
Succinimidyl-4-(N-maleimidomethyl)-cyclohexane-1-carboxy(6- LC SMCC
amidocaproate) N-(gamma-Maleimidobutryloxy)sulfosuccinimide ester
SULFO GMBS Succinimidyl-6-(beta-maleimidopropionamidohexanoate)
SMPH N-(kappa-Maleimidoundecanoyloxy)sulfosuccinimide ester SULFO
KMUS N-(gamma-Maleimidobutyrloxy)succinimide GMBS
Dimethyladipimidate hydrochloride DMA Dimethylpimelimidate
hydrochloride DMP Dimethylsuberimidate hydrochloride DMS
Methyl-p-hydroxybenzimidate hydrochloride, 98% MHBH(Wood's Reagent)
Amine Reactive Bis[sulfosuccinimidyl] suberate BS3
Bis[2-(succinimidooxycarbonyloxy)ethyl]sulfone BSOCOES
Disuccinimidyl glutarate DSG DSP (Lomant's Reagent)
1,5-Difluoro-2,4-dinitrobenzene DFDNB
Dithiobis[succinimidylpropionate DTBP
Bis-[b-(4-Azidosalicylamido)ethyl]disulfide BASED Sulfhydryl
Reactive BM[PEO].sub.3(1,8-bis-Maleimidotriethyleneglycol
BM[PEO].sub.3 BM[PEO].sub.4(1,11-bis-Maleimidotetraethyleneglycol
BM[PEO].sub.4 1,4-bis-Maleimidobutane BMB 1,4
bis-Maleimidyl-2,3-dihydroxybutane BMDB Bis-Maleimidohexane BMH
Bis-Maleimidoethane BMOE
1,4-Di-[3'-(2'-pyridyldithio)-propionamido]butane DPDPB
Dithio-bis-maleimidoethane DTME 1,6-Hexane-bis-vinylsulfone HBVS
p-Azidobenzoyl hydrazide ABH Amine-Sulfhydryl Reactive
N-[a-Maleimidoacetoxy]succinimide ester AMAS
N-[4-(p-Azidosalicylamido) butyl]-3'-(2'-pyridyldithio)propionamide
APDP N-[.beta.-Maleimidopropyloxy]succinimide ester BMPS
N-e-Maleimidocaproic acid EMCA N-e-Maleimidocaproyloxy]succinimide
ester EMCS N-[g-Maleimidobutyryloxy]succinimide ester GMBS
N-k-Maleimidoundecanoic acid KMUA
Succinimidyl-4-(N-Maleimidomethyl)cyclohexane-1-carboxy-(6- LC-SMCC
amidocaproate Succinimidyl
6-(3-[2-pyridyldithio]-propionamido)hexanoate LC-SPDP
m-Maleimidobenzoyl-N-hydroxysuccinimide ester MBS Succinimidyl
3-[bromoacetamido]propionate SBAP N-Succinimidyl iodoacetate SIA
N-Succinimidyl[4-iodoacetyl]aminobenzoate SIAB Succinimidyl
4-[N-maleimidomethyl]cyclohexane-1-carboxylate SMCC Succinimidyl
4-[p-maleimidophenyl]butyrate SMPB
Succinimidyl-6-[.beta.-maleimidopropionamido]hexanoate SMPH
4-Succinimidyloxycarbonyl-methyl-a-[2-pyridyldithio]toluene SMPT
N-Succinimidyl 3-[2-pyridyldithio]-propionamido SPDP
N-e-Maleimidocaproyloxy]sulfosuccinimide ester Sulfo-EMCS
N-[g-Maleimidobutyryloxy]sulfosuccinimide ester Sulfo-GMBS
N-[k-Maleimidoundecanoyloxy]sulfosuccinimide ester Sulfo-KMUS
4-Sulfosuccinimidyl-6-methyl-a-(2- Sulfo-LC-SMPT
pyridyldithio)toluamido]hexanoate Sulfosuccinimidyl
6-(3'-[2-pyridyldithio]-propionamido)hexanoate Sulfo-LC-SPDP
m-Maleimidobenzoyl-N-hydroxysulfosuccinimide ester Sulfo-MBS
N-Sulfosuccinimidyl[4-iodoacetyl]aminobenzoate Sulfo-SIAB
Sulfosuccinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate
Sulfo-SMCC Sulfosuccinimidyl-4-(P-Maleimidophenyl) Butyrate
Sulfo-SMPB Amino Groups N-5-Azido-2-nitrobenzoyloxysuccinimide
ANB-NOS Methyl N-succinimidyl adipate MSA
N-Hydroxysuccinimidyl-4-azidosalicylic acid NHS-ASA
N-Succinimidyl(4-azidophenyl)-1,3'-dithiopropionate SADP
Sulfosuccinimidyl 2-[7-amino-4-methylcoumarin-3-acetamido]ethyl-
SAED 1,3'dithiopropionate Sulfosuccinimidyl
2[m-azido-o-nitrobenzamido]-ethyl-1,3'- SAND dithiopropionate
N-Succinimidyl-6-[4'-azido-2'-nitrophenylamino] hexanoate SANPAH
Sulfosuccinimidyl-2-[p-azidosalicylamido]ethyl-1,3'-dithiopropionate
SASD Sulfosuccinimidyl-[perfluoroazidobenzamido]ethyl-1,3'- SFAD
dithiopropionate N-Hydroxysulfosuccinimidyl-4-azidobenzoate
Sulfo-HSAB Sulfosuccinimidyl[4-azidosalicylamido]-hexanoate
Sulfo-NHS-LC-ASA
N-Sulfosuccinimidyl(4-azidophenyl)-1,3'-dithiopropionate Sulfo-SADP
N-Sulfosuccinimidyl-6-[4'-azido-2'-nitrophenylamino] hexanoate
Sulfo-SANPAH p-Azidophenyl glyoxal monohydrate APG
N-.beta.-Maleimidopropionic acid BMPA Carbohydrate
Reactive-Photoreactive N-Succinimidyl-S-acetylthiopropionate SATP
Sulfhydryl-Carbohydrate Reactive 4-(4-N-Maleimidophenyl)butyric
acid hydrazide hydrochloride MPBH 3-(2-Pyridyldithio)propionyl
hydrazide PDPH Sulfhydryl-carbonyl (aldehyde)/carboxyl reactive
N-[.beta.-Maleimidopropionic acid]hydrazide.cndot.TFA BMPH
N-e-Maleimidocaproic acid]hydrazide EMCH N-[k-Maleimidoundecanoic
acid]hydrazide KMUH N-[p-Maleimidophenyl]isocyanate PMPI TFCS
Example 16
[1211] Fetal chromosomal abnormalities are determined by analyzing
SNPs wherein the maternal template DNA is homozygous and the
template DNA obtained from the plasma is heterozygous. Plasma that
is isolated from blood of a pregnant female contains both maternal
template DNA and fetal template DNA. Any number of SNP detection
methods can be used to analyze the maternal and plasma DNA. Any DNA
microarray may be used including but not limited to commercially
available and non-commercially available arrays.
[1212] A DNA microarray can be designed to contain SNPs located on
the chromosome or chromosomes of interest including but not limited
to a DNA microarray containing SNPs located on chromosomes 13, 18,
and 21, a DNA microarray containing SNPS located on chromosomes 13
and 18, a DNA microarray containing SNPS located on chromosomes 13
and 21, a DNA microarray containing SNPS located on chromosomes 18
and 21, a DNA microarray containing SNPS located on chromosomes 13,
18, 21, 15, 22, X, Y, a DNA microarray containing SNPS located on
each of the autosomal chromosomes and each of the sex chromosomes,
a DNA microarray containing SNPS located on chromosome 13, a DNA
microarray containing SNPS located on chromosome 18, a DNA
microarray containing SNPS located on chromosome 21, a DNA
microarray containing SNPS located on chromosome 15, a DNA
microarray containing SNPS located on chromosome 17, a DNA
microarray containing SNPS located on chromosome 22, a DNA
microarray containing SNPS located on a single chromosome, and a
DNA microarray containing SNPS located on multiple chromosomes.
[1213] In this example, SNPs are analyzed by GeneChip HuSNP Arrays
from Affymetrix, however any number of DNA arrays, including but
not limited to GeneChip arrays, GenFlex Tag arrays, Mapping 10K
Array, other Affymetrix arrays, and other DNA arrays can be
used.
Collection of Blood Samples
[1214] In accordance with an IRB approved study, blood samples are
collected from pregnant women after informed consent is granted.
Blood is collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284) and 0.225 ml of 10% neutral buffered solution containing
formaldehyde (4% w/v), is added to each tube, and each tube gently
is inverted. The tubes are stored at 4.degree. C. until ready for
processing.
[1215] Any number of agents that impede cell lysis or stabilize
cell membranes or cross-link cell membranes can be added to the
tubes including but not limited to formaldehyde, and derivatives of
formaldehyde, formal in, glutaraldehyde, and derivatives of
glutaraldehyde, crosslinkers, primary amine reactive crosslinkers,
sulfhydryl reactive crosslinkers, sulfhydryl addition or disulfide
reduction, carbohydrate reactive crosslinkers, carboxyl reactive
crosslinkers, photoreactive crosslinkers, cleavable crosslinkers,
AEDP, APG, BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH,
BMOE, BS3, BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS,
DST, DTBP, DTME, DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST,
Sulfo-EGS or compounds listed in Table XXIII. Any concentration of
agent that stabilizes cell membranes, impedes cell lysis or
cross-link cell membranes can be added. In a preferred embodiment,
the agent that stabilizes cell membranes, impedes cell lysis, or
cross-links cell membranes is added at a concentration that does
not impede or hinder subsequent reactions.
[1216] An agent that stabilizes cell membranes may be added to the
maternal blood sample to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, mexitil, dilantin, sodium hyaluronate, or polaxamer
188.
Isolation of Plasma and Maternal Cells
[1217] The blood is stored at 4.degree. C. until processing. The
tubes are spun at 1000 rpm for ten minutes in a centrifuge with
braking power set at zero. The tubes are spun a second time at 1000
rpm for ten minutes. The supernatant (the plasma) of each sample is
transferred to a new tube and spun at 3000 rpm for ten minutes with
the brake set at zero. The supernatant is transferred to a new tube
and stored at -80.degree. C. Approximately two milliliters of the
"buffy coat," which contains maternal cells, is placed into a
separate tube and stored at -80.degree. C.
Isolation of DNA
[1218] DNA is isolated from the plasma sample using the Qiagen Midi
Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA is eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also is used to isolate DNA from the maternal cells
contained in the "buffy coat."
Identification of Homozygous Maternal SNPs
HuSNP Assay
[1219] The HuSNP assay is done as described by K. Lindblad-Toh et
al. (Nature Biotechnology, Vol. 18, 1001-1005). The GeneChip.RTM.
HuSNP.TM. Array is thought to enable whole genome surveys by
simultaneously tracking nearly 1,500 SNPs dispersed throughout the
genome. In this example, HuSNP array is used as a representative
Affymetrix array, and is not meant to limit the use of other arrays
including but not limited to GeneChip CYP450, and Affymetrix custom
arrays that are designed to meet specific user requirements.
PCR Amplification
[1220] Maternal DNA is assayed according to the HuSNP protocol
supplied by Affymetrix. Inc. For each sample, 24 pools of primer
pairs (50-100 loci/pool at 50 nM each) are mixed with 5 ng of
maternal DNA, 5 mM MgCl.sub.2, 0.5 mM dNTPs, 1.25 U Amplitaq Gold
(PE Biosystems, Foster City, Calif.), and the supplied buffer in
12.5 .mu.l per pool. Samples are denatured for 5 min at 95.degree.
C. followed by 30 cycles of 95.degree. C. for 30 s, 52.degree.
C.+0.2.degree. C./cycle for 55 s, and 72.degree. C. for 30 s; 5
cycles of 95.degree. C. for 30 s, 58.degree. C. for 55 s, and
72.degree. C. for 30 s and a final extension of 72.degree. C. for 7
min. A 1:1000 dilution of each pool is made by adding 1 .mu.l of
the amplification product to 999 .mu.l of ddH20. After, 2.5 .mu.l
of the 1:1000 dilution is transferred to a new plate and amplified
with 0.8 .mu.M biotinylated T7 and 0.8 .mu.M biotinylated T3
primers, 4 mM MgCl2, 0.4 mM dNTPs, 2.5 U Taq and the supplied
buffer in 25 .mu.l for 8 min at 95.degree. C., followed by 40
cycles of 95.degree. C. for 30 s, 55.degree. C. for 90 s, and
72.degree. C. for 30 s, and a final extension of 72.degree. C. for
7 min. Then 1.5 .mu.l from each pool is tested for amplification on
a 3% agarose gel. For each sample, the remainder of each the 24
pools is mixed and loaded on a Microcon-10 spin column (Amicon
Bioseparations, Bedford, Mass.). Samples are concentrated by
spinning the column for 20 min at 13,000g at room temperature and
are eluted by inverting the column and centrifuging for 3 min at
3,000g. Volumes are adjusted to 60 .mu.l
[1221] A custom array can be designed using only the SNPs that are
of interest. For example, a custom array may be designed that
contains SNPs that are located on chromosomes 1, 13, 21, 18, 15, X,
and Y.
[1222] Additionally, any number of SNPs can be amplified including
SNPs located on any human chromosome including chromosome 1, 2, 3,
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21,
22, X or Y. Two representative SNPs on chromosome 13 and two
representative SNPs on chromosome 21 are chosen. The genomic
location and sequence of SNPs may be found at the SNP consortium
(http://snp.cshl.org). If these SNPs are not present on the array,
different SNPs can be chosen.
[1223] SNP TSC0466917 (C/G), which is located on chromosome 13, is
amplified using the following primers:
TABLE-US-00238 Upstream Primer: 5' CCAGCTGGTAGAACTT 3' (SEQ ID NO:
629) Downstream Primer: 5' CCCAATAGACCTATAG 3' (SEQ ID NO: 630)
[1224] SNP TSC1172576 (T/A), which is located on chromosome 13, is
amplified using the following primers:
TABLE-US-00239 Upstream Primer: 5' TAGCAGAATCTCTCAT 3' (SEQ ID NO:
631) Downstream Primer: 5' AGAGTATCTCATTTGTT 3' (SEQ ID NO:
632)
[1225] SNP TSC0271628 (A/G), which is located on chromosome 21, is
amplified using the following primers:
TABLE-US-00240 Upstream Primer: 5' AGGAAATTGTGAAGTA 3' (SEQ ID NO:
633) Downstream Primer: 5' TAACTCACTCACTATC 3' (SEQ ID NO: 634)
[1226] SNP TSC0069805 (C/T), which is located on chromosome 21, is
amplified using the following primers:
TABLE-US-00241 Upstream Primer: 5' CTGCTGAGTCATAGTC 3' (SEQ ID NO:
635) Downstream Primer: 5' TGTTCTTTGAATCAAC 3' (SEQ ID NO: 636)
Hybridization to GeneChip Probe Arrays, Washing and Staining
[1227] 5-30 .mu.l of the sample (depending on the intensity of the
chip lot) is diluted in 3 M tetramethylammonium chloride (TMACl), 2
mM control oligonucleotide B1 (supplied by Affymetrix),
5.times.Denhardt's solution, 100 .mu.g/ml herring sperm DNA, 5 mM
EDTA pH 8.0, 10 mM Tris pH7.8, and 0.01% Tween 20 in a volume of
135 .mu.l and is denatured for 10 min at 95.degree. C. After two
minutes on ice, the samples are loaded into HuSNP chips and
hybridized for 16 h at 44.degree. C. and 40 r.p.m.
[1228] Each chip is washed and stained on the Affymetrix fluidics.
Chips are washed for two cycle's of two mixes with 6.times.SSPET
(Bio Whitaker, Walkersville, Md.) (6.times.SSPE (sodium chloride,
sodium phosphate, sodium EDTA)+0.01% Triton-X-100) at 25.degree.
C., and for six cycles of five mixes with 4.times.SSPET
(4.times.SSPE+0.01% Triton X-100) at 35.degree. C. Chips are
stained for 30 min at 25.degree. C. with 50 .mu.g/ml
streptavidin-phycoerhthrin and 0.25 mg/ml biotinylated
anti-streptavidin antibody in 6.times.SSPE, 1.times.Denhardt's
solution, and 0.01% Tween 20 in a volume of 500 .mu.l. The chip is
filled with 6.times.SSPET following six washes of four mixes with
6.times.SSPET at 25.degree. C.
[1229] After the hybridization, washing, and staining procedures,
the HuSNP probe arrays are scanned using the HP GeneArray Scanner
(HuSNP Mapping Assay Manual Affymetrix P/N 700308).
Scanning
[1230] The HuSNP probe arrays are scanned using the HP GeneArray
Scanner according to the HuSNP Mapping Assay Manual (Affymetrix P/N
700308). Other scanners may be used including but not limited to
the AlphaArray.TM. Reader. Genotype calls are made automatically
from the collected hybridization signal intensities by the
Affymetrix Microarray Suite version 5.0 software. Each allele of a
SNP is represented by four or five complementary probes with
different locations of the SNP base position within the
20-nucleotide probes. Each of these probes, in turn, is paired with
a probe of the same sequence except for a central mismatch at or
near a SNP position, intended to correct the fluorescence value for
non-specific binding to the probe.
[1231] Each SNP is genotyped. SNPs located on chromosomes 13 and
21, wherein the maternal DNA is homozygous, are analyzed with the
DNA isolated from the plasma.
Analysis of DNA Isolated from Maternal Plasma
[1232] After the maternal DNA is analyzed and homozygous SNPs are
identified, these SNPs are analyzed with the DNA isolated from the
plasma. A low copy number of fetal genomes typically is present in
the maternal plasma. To increase the copy number of the loci of
interest, which are the SNPs at which the maternal DNA is
homozygous, primers are designed to anneal at approximately 130
bases upstream and 130 bases downstream of each loci of interest.
This is done to reduce statistical sampling error that can occur
when working with a low number of genomes, which can influence the
ratio of one allele to another (see Example 11).
Design of Multiplex Primers
[1233] The primers are 12 bases in length. However, primers of any
length can be used including but not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers are designed to anneal to both the sense strand and
the antisense strand.
[1234] The maternal homozygous SNPs vary from sample to sample so
defined sequences are not provided here. Primers are designed to
anneal about 130 bases upstream and downstream of the maternal
homozygous SNPs. The primers are designed to terminate at the 3'
end in the dinucleotide "AA" to reduce the formation of
primer-dimers. However, the primers can be designed to end in any
of the four nucleotides and in any combination of the four
nucleotides.
Multiplex PCR
[1235] Regions upstream and downstream of the maternal homozygous
SNPs are amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
uses primers that anneal approximately 130 bases upstream and
downstream of each loci of interest. The primers are mixed together
and are used in a single reaction to amplify the template DNA. This
reaction is done to increase the number of copies of the loci of
interest, which eliminates error generated from a low number of
genomes.
[1236] For increased specificity, a "hot-start" PCR reaction is
used. PCR reactions are performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction is optimized for each locus of
interest. In this example, the 20 .mu.l of plasma template DNA is
used.
[1237] Two microliters of each forward and reverse primer, at
concentrations of 5 mM are pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix is used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.A of primer mix, and 25 .mu.l
of HotStar Taq, Twenty-five cycles of PCR are performed. The
following PCR conditions are used: [1238] (1) 95.degree. C. for 15
minutes; [1239] (2) 95.degree. C. for 30 second; [1240] (3)
4.degree. C. for 30 seconds; [1241] (4) 37.degree. C. for 30
seconds; [1242] (5) Repeat steps 2-4 twenty-four (24) times; [1243]
(6) 72.degree. C. for 10 minutes.
[1244] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
[1245] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
[1246] It is important to ensure that the region amplified (done to
increase the copy number of the fetal loci of interest) contains
annealing sequences for the primers that are used with the CodeLink
assay. Upon purchase of the CodeLink array, each SNP and the
primers used to amplify each SNP can be identified. With this
knowledge, the multiplex primers are designed to encompass
annealing regions for the primers in the HuSNP Array.
Purification of Fragment of Interest
[1247] The unused primers, and nucleotides are removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions are performed following the
manufacturer's instructions supplied with the columns. The DNA is
eluted in 100 .mu.l of sterile water. 5 .mu.l of each amplified
loci is mixed together
CodeLink Assay, Washing, Staining and Scanning
[1248] The pooled DNA is assayed with the CodeLink Array as
described above. Washing, staining, and scanning procedures are as
described above.
[1249] Each SNP is genotyped. SNPs located on chromosomes 13 and
21, wherein the maternal DNA is homozygous, and DNA isolated from
the plasma is heterozygous are quantitated.
Quantification
[1250] The intensity of the signal for each allele at a
heterozygous is SNP is quantitated. As discussed above, the
expected ratio of allele 1 to allele 2 can be used to determine the
presence or absence of a chromosomal abnormality. If the maternal
genome is homozygous at SNP X (A/A), and the plasma DNA is
heterozygous at SNP X (A/G), then the G represents the distinct
fetal signal. The ratio of G:A depends on the percentage of fetal
DNA present in the maternal blood.
[1251] For example, if the sample contains 50% fetal DNA, then the
expected ratio is 0.33 (1 fetal G allele/(2 maternal A alleles+1
fetal A allele)). This ratio should be constant for all chromosomes
that are present in two copies. The ratio that is obtained for SNPs
on chromosome 13 should be the same as the ratio that is obtained
for chromosome 21.
[1252] However, if the fetal genome contains an additional copy of
chromosome 21, then the ratio for this chromosome will deviate from
the expected ratio. The expected ratio for a Trisomy condition with
50% fetal DNA in the maternal blood is 0.25. Thus, by analyzing
SNPs wherein the maternal genome is homozygous, and the DNA that is
isolated from the plasma is heterozygous, fetal chromosomal
abnormalities can be detected.
[1253] This example explained the use of CodeLink Arrays, but it
not intended to limit the use of arrays. Any DNA array may be used
including but not limited to the DNA arrays listed in Table XXIII,
or DNA arrays available from any of the companies listed in Table
XXIV.
Example 18
[1254] Fetal chromosomal abnormalities are determined by analyzing
SNPs wherein the maternal template DNA is homozygous and the
template DNA obtained from the plasma is heterozygous. Plasma that
is isolated from blood of a pregnant female contains both maternal
template DNA and fetal template DNA. Any number of SNP detection
methods can be used to analyze the maternal and plasma DNA. In this
example, SNPs are analyzed using Illumina's BeadArray.TM. platform,
available form Illumina in San Diego, Calif.
Collection of Blood Samples
[1255] In accordance with an IRB approved study, blood samples are
collected from pregnant women after informed consent is granted.
Blood is collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284) and 0.225 ml of 10% neutral buffered solution containing
formaldehyde (4% w/v), is added to each tube, and each tube gently
is inverted. The tubes are stored at 4.degree. C. until ready for
processing.
[1256] Any number of agents that impede cell lysis or stabilize
cell membranes can be added to the tubes including but not limited
to formaldehyde, and derivatives of formaldehyde, formalin,
glutaraldehyde, and derivatives of glutaraldehyde, crosslinkers,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP,
DTME, DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS or
compounds listed in Table XXIII. Any concentration of agent that
stabilizes cell membranes or impedes cell lysis can be added. In a
preferred embodiment, the agent that stabilizes cell membranes or
impedes cell lysis is added at a concentration that does not impede
or hinder subsequent reactions.
[1257] An agent that stabilizes cell membranes may be added to the
maternal blood sample to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, dilantin, sodium hyaluronate, or polaxamer 188.
Isolation of Plasma and Maternal Cells
[1258] The blood is stored at 4.degree. C. until processing. The
tubes are spun at 1000 rpm for ten minutes in a centrifuge with
braking power set at zero. The tubes are spun a second time at 1000
rpm for ten minutes. The supernatant (the plasma) of each sample is
transferred to a new tube and spun at 3000 rpm for ten minutes with
the brake set at zero. The supernatant is transferred to a new tube
and stored at -80.degree. C. Approximately two milliliters of the
"buffy coat," which contains maternal cells, is placed into a
separate tube and stored at -80.degree. C.
Isolation of DNA
[1259] DNA is isolated from the plasma sample using the Qiagen Midi
Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA is eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also is used to isolate DNA from the maternal cells
contained in the "buffy coat."
Identification of Homozygous Maternal SNPs
[1260] Illumina's BeadArray.TM. technology consists of a fiber
optic based array system that allegedly allows miniaturized,
very-high throughput genetic analysis, Illumina's 96-bundle Sentrix
Array.TM. allegedly enable parallel processing of nearly 150,000
SNPs.
[1261] Fiber bundles are manufactured to contain nearly 50,000
individual, light transmitting fiber strands. Each fiber bundle is
converted into an array by first chemically etching a microscopic
well at the end of each fiber strand within a bundle, which creates
up to 50,000 discrete microscopic wells per bundle.
[1262] In a separate process, sensors are created by affixing a
specific type of molecule to the beads, each bead approximately 3
microns in diameter. For SNP analysis, a particular DNA sequence is
attached to each bead in a batch. Illumina states that hundreds of
thousands of molecules of the same type coat each bead. Batches of
coated beads are combined to form a pool specific to the type of
array desired. For SNP analysis, the array pool allegedly uses DNA
sequences that do not cross hybridize with themselves or with known
genomic DNA.
[1263] Next, the self-assembled array is created. By dipping
bundles into a pre-mixed bead pool, the coated beads self-assemble
individually, one bead per well, on the end of each fiber in the
bundle to create the array. In Illumina's SNP genotyping array, the
bead pool consists of up to 1500 sequences, which self assemble in
each bundle of 50,000 fibers to create an array with an
approximately thirty-fold redundancy.
[1264] The BeadArray bundles are assembled into a matrixed device,
which is called the Array of Arrays.TM. platform, where each fiber
bundle of the larger array matches a well of a standardized
microtiter plate.
[1265] Following array assembly, a decoding process is used to
determine the bead type that resides in each fiber core. The DNA
molecules are synthesized using the Oligator.TM. custom DNA
synthesis technology.
[1266] Illumina's SNP genotyping service using the BeadArray
technology and other technologies that have sprung from the
BeadArray technology are provided at Illumina's facilities or at
facilities that have received a license to the BeadArray
technology.
[1267] Maternal DNA samples are analyzed using BeadArrays that
contain oligonucleotide probes to SNPs. The oligonucleotide probes
can be for SNPs located on any chromosome including human
chromosome 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
17, 18, 19, 20, 21, 22, X, and Y. The BeadArrays are analyzed to
identify SNPs, wherein the maternal template DNA is homozygous. The
identified homozygous SNPs are then analyzed using the DNA isolated
from the maternal plasma.
Analysis of DNA Isolated from Maternal Plasma
[1268] After the maternal DNA is analyzed and homozygous SNPs are
identified, these SNPs are analyzed with the DNA isolated from the
plasma. A low copy number of fetal genomes typically exists in the
maternal plasma. To increase the copy number of the loci of
interest, which are the SNPs at which the maternal DNA is
homozygous, primers are designed to anneal at approximately 130
bases upstream and 130 bases downstream of each loci of interest.
This is done to reduce statistical sampling error that can occur
when working with a low number of genomes; which can influence the
ratio of one allele to another (see Example 11).
Design of Multiplex Primers
[1269] The primers are 12 bases in length. However, primers of any
length can be used including but not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers are designed to anneal to both the sense strand and
the antisense strand.
[1270] The maternal homozygous SNPs vary from sample to sample so
defined sequences are not provided here. Primers are designed to
anneal about 130 bases upstream and downstream of the maternal
homozygous SNPs. The primers are designed to terminate at the 3'
end in the dinucleotide "AA" to reduce the formation of
primer-dimers. However, the primers can be designed to end in any
of the four nucleotides and in any combination of the four
nucleotides.
Multiplex PCR
[1271] Regions upstream and downstream of the maternal homozygous
SNPs are amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
uses primers that anneal approximately 130 bases upstream and
downstream of each loci of interest. The primers are mixed together
and are used in a single reaction to amplify the template DNA. This
reaction is done to increase the number of copies of the loci of
interest, which eliminates error generated from a low number of
genomes.
[1272] For increased specificity, a "hot-start" PCR reaction is
used. PCR reactions are performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction is optimized for each locus of
interest. In this example, the 20 .mu.l of plasma template DNA is
used.
[1273] Two microliters of each forward and reverse primer, at
concentrations of 5 mM are pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix is used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.l of primer mix, and 25 .mu.l
of HotStar Taq. Twenty-five cycles of PCR are performed. The
following PCR conditions are used: [1274] (1) 95.degree. C. for 15
minutes; [1275] (2) 95.degree. C. for 30 second; [1276] (3)
4.degree. C. for 30 seconds; [1277] (4) 37.degree. C. for 30
seconds; [1278] (5) Repeat steps 2-4 twenty-four (24) times; [1279]
(6) 72.degree. C. for 10 minutes.
[1280] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
[1281] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
[1282] It is important to ensure that the region amplified contains
annealing sequences for the oligonucleotide probes in the
BeadArray. Upon purchase of the BeadArray service, each SNP and the
primers used to analyze each SNP are identified. With this
knowledge, the multiplex primers are designed to encompass
annealing regions for the primers in the BeadArray.
Purification of Fragment of Interest
[1283] The unused primers, and nucleotides are removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions are performed following the
manufacturer's instructions supplied with the columns. The DNA is
eluted in 100 .mu.l of sterile water. 5 .mu.l of each amplified
loci is mixed together
BeadArray Technology
[1284] The pooled DNA is assayed with the BeadArray as described
above. Each SNP is genotyped. SNPs located on chromosomes 13 and
21, wherein the maternal DNA is homozygous, and DNA isolated from
the plasma is heterozygous are quantitated.
Quantification
[1285] The intensity of the signal for each allele at a
heterozygous is SNP is quantitated. As discussed above, the
expected ratio of allele 1 to allele 2 can be used to determine the
presence or absence of a chromosomal abnormality. If the maternal
genome is homozygous at SNP X (A/A), and the plasma DNA is
heterozygous at SNP X (A/G), then the G represents the distinct
fetal signal. The ratio of G:A depends on the percentage of fetal
DNA present in the maternal blood.
[1286] For example, if the sample contains 50% fetal DNA, then the
expected ratio is 0.33 (1 fetal G allele/(2 maternal A alleles+1
fetal A allele)). This ratio should be constant for all chromosomes
that are present in two copies. The ratio that is obtained for SNPs
on chromosome 13 should be the same as the ratio that is obtained
for chromosome 21.
[1287] However, if the fetal genome contains an additional copy of
chromosome 21, then the ratio for this chromosome will deviate from
the expected ratio. The expected ratio for a Trisomy condition with
50% fetal DNA in the maternal blood is 0.25. Thus, by analyzing
SNPs wherein the maternal genome is homozygous, and the DNA that is
isolated from the plasma is heterozygous, fetal chromosomal
abnormalities can be detected.
[1288] This example explained the use of Illumines BeadArray
Technology, but it not intended to limit the use of arrays. Any DNA
array may be used including but not limited to the DNA arrays
listed in Table XXIII, or DNA arrays available from any of the
companies listed in Table XXIV.
Example 19
[1289] Fetal chromosomal abnormalities are determined by analyzing
SNPs wherein the maternal template DNA is homozygous and the
template DNA obtained from the plasma is heterozygous. Plasma that
is isolated from blood of a pregnant female contains both maternal
template DNA and fetal template DNA. Any number of SNP detection
methods can be used to analyze the maternal and plasma DNA. In this
example, SNPs are analyzed using Sequenom's MassArray.TM. System,
which uses Sequenom's homogenous MassCleave.TM. (hMC) method.
Collection of Blood Samples
[1290] In accordance with an IRB approved study, blood samples are
collected from pregnant women after informed consent is granted.
Blood is collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284) and 0.225 ml of 10% neutral buffered solution containing
formaldehyde (4% w/v), is added to each tube, and each tube gently
is inverted. The tubes are stored at 4.degree. C. until ready for
processing.
[1291] Any number of agents that impede cell lysis or stabilize
cell membranes can be added to the tubes including but not limited
to formaldehyde, and derivatives of formaldehyde, formalin,
glutaraldehyde, and derivatives of glutaraldehyde, crosslinkers,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP,
DTME, DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS or
compounds listed in Table XXIII. Any concentration of agent that
stabilizes cell membranes or impedes cell lysis can be added. In a
preferred embodiment, the agent that stabilizes cell membranes or
impedes cell lysis is added at a concentration that does not impede
or hinder subsequent reactions.
[1292] An agent that stabilizes cell membranes may be added to the
maternal blood sample to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, mexitil, dilantin, sodium hyaluronate, or polaxamer
188.
Isolation of Plasma and Maternal Cells
[1293] The blood is stored at 4.degree. C. until processing. The
tubes are spun at 1000 rpm for ten minutes in a centrifuge with
braking power set at zero. The tubes are spun a second time at 1000
rpm for ten minutes. The supernatant (the plasma) of each sample is
transferred to a new tube and spun at 3000 rpm for ten minutes with
the brake set at zero. The supernatant is transferred to a new tube
and stored at -80.degree. C. Approximately two milliliters of the
"buffy coat," which contains maternal cells, is placed into a
separate tube and stored at -80.degree. C.
Isolation of DNA
[1294] DNA is isolated from the plasma sample using the Qiagen Midi
Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA is eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also is used to isolate DNA from the maternal cells
contained in the "buffy coat."
Identification of Homozygous Maternal SNPs
Targeted SNP Discovery: hMC Method
[1295] Sequenom's hMC method uses nucleotide base-specific cleavage
for genotyping. The cleaved fragments are measured using MALDI-TOF
to generate a characteristic peak signal, based on the mass of each
fragment, for any particular sequence.
Primer Design
[1296] Four primers are needed for the two PCR reactions (one
forward reaction and one reverse reaction). The recommended size
range for PCR amplicons is 300-700 base pairs. The primers contain
a T-7 promoter tagged forward or reverse primer to obtain an
appropriate product for in vitro transcription. An 8 base insert is
included to prevent abortive cycling. The primer that lacks the T-7
promoter contains a 10-mer tag in order to balance the primers.
[1297] The primers for one SNP are provided below. SNP TSC1172576
(T/A), which is located on chromosome 13, is amplified using the
following primers for the forward reaction:
[1298] Forward Reaction:
TABLE-US-00242 Upstream Primer: (SEQ ID NO: 637) 5'
CAGTAATACGACTCACTATAGGGGTCAGGATTAGCAGAATCTCTC AT 3' Downstream
Primer: (SEQ ID NO: 638) 5' GCATTCTATGAGAGTATCTCATTTGTT 3'
[1299] Reverse Reaction:
TABLE-US-00243 Upstream Primer (SEQ ID NO: 639) 5'
CAGTAATACGACTCACTATAGGGGTCAGGAAGAGTATCTCATTTGT T 3' Downstream
Primer (SEQ ID NO: 640) 5' GCATTCTATGTAGCAGAATCTCTCAT 3'
[1300] The sequence of the T-7 promoter are in italics, the 8 base
insert is underlined, the 10-base balancing sequence is double
underlined, and the gene specific sequences are unmodified.
PCR Amplification
[1301] Five nanograms of DNA is amplified in a 5 .mu.l volume using
a 384-microtiter format. The following PCR conditions are used:
[1302] 1) 94.degree. C. for 15 minutes;
[1303] 2) 94.degree. C. for 20 seconds;
[1304] 3) 62.degree. C. for 30 seconds;
[1305] 4) 72.degree. C. for 1 minute;
[1306] 5) Repeat steps 2-4 44 times; and
[1307] 6) 72.degree. C. for 3 minutes.
Dephosphorylation
[1308] Shrimp Alkaline Phosphatase (SAP) (2 .mu.l) is added to each
5 .mu.l PCR reaction to dephosphorylate unincorporated dNTPs from
the PCR reaction. The plates are incubated at 37.degree. C. for 20
minutes. Then, the plates are incubated at 85.degree. C. for 5
minutes.
In Vitro Transcription
[1309] For each transcription reaction, 2 .mu.l of transcription
cocktail and 2 .mu.l of PCR/SAP sample are needed. Add 2 .mu.l of
transcription cocktail and 2 .mu.l of PCR/SAP sample to a new
microtiter plate. The plates are incubated at 37.degree. C. for two
hours. For detailed information regarding these protocols see the
"Processing homogeneous MassCLEAVE Reactions" chapter in the
MassARRAY Liquid Handler SNP Discovery User's Guide for
instructions, which is fully incorporated herein by reference,
RNase A Cleavage
[1310] RNase A cocktail (2.5 .mu.l) is added to each reaction (T
cleavage and C cleavage). The plates are incubated at 37.degree. C.
for one hour.
[1311] Depending on the nucleotide at the SNP site, various
fragments of different weights are generated. For example, the DNA
sequence surrounding SNP TSC1172576, which is located on chromosome
13, is as follows:
TABLE-US-00244 5' CCGCATA T/A CTCAGCACA 3' (SEQ ID NO: 641)
3'GGCGTAT A/T GAGTCGTGT 5' (SEQ ID NO: 642)
[1312] After PCR, in vitro transcription, and base-specific
cleavage, the following fragments for each allele are
generated:
TABLE-US-00245 T allele A allele Products of forward 1) TATCTCA 1)
CTCAGC Transcription 2) ATC 2) AAC Products of reverse 1) AGTTA 1)
AGATA transcription 2) GTTATGC 2) ATATGCG
[1313] For the ATC and AAC fragments, the weight difference between
T and A is used to determine the genotype at SNP TSC1172576.
Likewise, the weight difference between T and A in fragments AGTTA
and AGATA is used to determine the genotype at SNP TSC1172576.
Sample Conditioning
[1314] Double distilled water (20 .mu.l) is added to each sample
within the 384-well plate. Clean Resin (6 mg) is added to each
well. The plate is rotated for 10 minutes, followed by a
centrifugation at 3200.times.g. It is recommended that water always
be added before the Clean Resin.
Sample Transfer
[1315] The hMC reaction product (10-15 .mu.l is dispensed onto a
384 element SpectroCHIP.RTM.. For further information, see the
"Dispensing MassCLEAVE Reaction Products onto SpectroCHIPs" chapter
in the MassARRAY Nanodispenser SNP Discovery User's Guide for
instructions.
Sample Analysis
[1316] Spectra from the four cleavage reactions is acquired using
the MassARRAY.TM. system. For further instructions, see the
"Acquiring Spectra" chapter in the MassARRAY Discovery RT Software
User's guide for instructions on acquiring spectra from
SpectroCHIPS.RTM..
SNP Analysis
[1317] The results are analyzed using the SNP Discovery Analysis
software. For further instructions, see the "Analyzing SNPs"
chapter in the MassARRAY Discovery RT Software User's Guide for
instructions on using the SNP Discovery Analysis software.
Components that are useful for the MassARRAY procedure include
MassARRAY.TM. Analyzer (part number 004500), MassARRAY.TM.
Discovery RT Software version 1.2 (part number 11434),
MassARRAY.TM. SNP Discovery Starter Kit (part number 10027), and
Liquid Handler SNP Discovery Methods and Macros (part number
11433).
[1318] The SpectroCHIP array is used to genotype the maternal DNA
following the manufacturer's recommended protocols and procedures,
which are made available after purchase of the SpectroCHIP array.
SNPs at which the maternal DNA is homozygous are used to analyze
the DNA isolated from the maternal plasma.
Analysis of DNA Isolated from Maternal Plasma
[1319] After the maternal DNA is analyzed and homozygous SNPs are
identified, these SNPs are analyzed with the DNA isolated from the
plasma. A low copy number of fetal genomes typically exists in the
maternal plasma. To increase the copy number of the loci of
interest, which are the SNPs at which the maternal DNA is
homozygous, primers are designed to anneal at approximately 130
bases upstream and 130 bases downstream of each loci of interest.
This is done to reduce statistical sampling error that can occur
when working with a low number of genomes, which can influence the
ratio of one allele to another (see Example 11).
Design of Multiplex Primers
[1320] The primers are 12 bases in length. However, primers of any
length can be used including but not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers are designed to anneal to both the sense strand and
the antisense strand.
[1321] The maternal homozygous SNPs vary from sample to sample so
defined sequences are not provided here. Primers are designed to
anneal about 130 bases upstream and downstream of the maternal
homozygous SNPs. The primers are designed to terminate at the 3'
end in the dinucleotide "AA" to reduce the formation of
primer-dimers. However, the primers can be designed to end in any
of the four nucleotides and in any combination of the four
nucleotides.
Multiplex PCR
[1322] Regions upstream and downstream of the maternal homozygous
SNPs are amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
uses primers that anneal approximately 130 bases upstream and
downstream of each loci of interest. The primers are mixed together
and are used in a single reaction to amplify the template DNA. This
reaction is done to increase the number of copies of the loci of
interest, which eliminates error generated from a low number of
genomes.
[1323] For increased specificity, a "hot-start" PCR reaction is
used. PCR reactions are performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction is optimized for each locus of
interest. In this example, the 20 .mu.l of plasma template DNA is
used.
[1324] Two microliters of each forward and reverse primer, at
concentrations of 5 mM are pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix is used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.l of primer mix, and 25 .mu.l
of HotStar Taq. Twenty-five cycles of PCR are performed. The
following PCR conditions are used: [1325] (1) 95.degree. C. for 15
minutes; [1326] (2) 95.degree. C. for 30 second; [1327] (3)
4.degree. C. for 30 seconds; [1328] (4) 37.degree. C. for 30
seconds; [1329] (5) Repeat steps 2-4 twenty-four (24) times; [1330]
(6) 72.degree. C. for 10 minutes.
[1331] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
[1332] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
[1333] It is important to ensure that the region amplified contains
annealing sequences for the PCR primers in the targeted SNP
discovery, hMC method
Purification of Fragment of Interest
[1334] The unused primers, and nucleotides are removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions are performed following the
manufacturer's instructions supplied with the columns. The DNA is
eluted in 100 .mu.l of sterile water. 5 .mu.l of each amplified
loci is mixed together
Targeted SNP Discovery: hMC Method
[1335] The pooled DNA is assayed with the INC method as described
above. Each SNP is genotyped. SNPs located on chromosomes 13 and
21, wherein the maternal DNA is homozygous, and DNA isolated from
the plasma is heterozygous are quantitated. However, SNPs located
on other chromosomes can also be quantitated if so desired.
Quantification
[1336] The intensity of each peak, wherein each peak corresponds to
a DNA fragment with a specific molecular weight, is quantitated. As
discussed above, the expected ratio of allele 1 to allele 2 is used
to determine the presence or absence of a chromosomal abnormality.
If the maternal genome is homozygous at SNP X (NA), and the plasma
DNA is heterozygous at SNP X (A/G), then the G represents the
distinct fetal signal.
[1337] There will be some fragments that differ in molecular due to
the presence of the G nucleotide at SNP X in the fetal genome. The
intensity of the peak with the A nucleotide is quantitated and the
intensity of the peak that corresponds to fragments with the G
nucleotide is quantitated. The ratio of G:A depends on the
percentage of fetal DNA present in the maternal blood.
[1338] For example, if the sample contains 50% fetal. DNA, then the
expected ratio is 0.33 (1 fetal U allele/(2 maternal A alleles+1
fetal A allele)). This ratio should be constant for all chromosomes
that are present in two copies. The ratio that is obtained for SNPs
on chromosome 13 should be the same as the ratio that is obtained
for chromosome 21.
[1339] However, if the fetal genome contains an additional copy of
chromosome 21, then the ratio for this chromosome will deviate from
the expected ratio. The expected ratio for a Trisomy condition with
50% fetal DNA in the maternal blood is 0.25. Thus, by analyzing
SNPs wherein the maternal genome is homozygous, and the DNA that is
isolated from the plasma is heterozygous, fetal chromosomal
abnormalities can be detected.
[1340] This example explained the use of Sequenom's hMC method, but
it not intended to limit the use of other mass spectrometry
techniques.
Example 20
[1341] Fetal chromosomal abnormalities are determined by analyzing
SNPs wherein the maternal template DNA is homozygous and the
template DNA obtained from the plasma is heterozygous. Plasma that
is isolated from blood of a pregnant female contains both maternal
template DNA and fetal template DNA. Any number of SNP detection
methods can be used to analyze the maternal and plasma DNA. In this
example, SNPs are analyzed using Sequenom's MassArray.TM.
Homogenous MassEXTENDT.TM. (hME) Assay.
Collection of Blood Samples
[1342] In accordance with an IRB approved study, blood samples are
collected from pregnant women after informed consent is granted.
Blood is collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284) and 0.225 ml of 10% neutral buffered solution containing
formaldehyde (4% w/v), is added to each tube, and each tube gently
is inverted. The tubes are stored at 4.degree. C. until ready for
processing.
[1343] Any number of agents that impede cell lysis or stabilize
cell membranes can be added to the tubes including but not limited
to formaldehyde, and derivatives of formaldehyde, formalin,
glutaraldehyde, and derivatives of glutaraldehyde, crosslinkers,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP,
DTME, DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfa-EGS or
compounds listed in Table XXIII. Any concentration of agent that
stabilizes cell membranes or impedes cell lysis can be added. In a
preferred embodiment, the agent that stabilizes cell membranes or
impedes cell lysis is added at a concentration that does not impede
or hinder subsequent reactions.
[1344] An agent that stabilizes cell membranes may be added to the
maternal blood sample to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, mexitil, dilantin, sodium hyaluronate, or polaxamer
188.
Isolation of Plasma and Maternal Cells
[1345] The blood is stored at 4.degree. C. until processing. The
tubes are spun at 1000 rpm for ten minutes in a centrifuge with
braking power set at zero. The tubes are spun a second time at 1000
rpm for ten minutes. The supernatant (the plasma) of each sample is
transferred to a new tube and spun at 3000 rpm for ten minutes with
the brake set at zero. The supernatant is transferred to a new tube
and stored at -80.degree. C. Approximately two milliliters of the
"buffy coat," which contains maternal cells, is placed into a
separate tube and stored at -80.degree. C.
Isolation of DNA
[1346] DNA is isolated from the plasma sample using the Qiagen Midi
Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA is eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also is used to isolate DNA from the maternal cells
contained in the "buffy coat."
Identification of Homozygous Maternal SNPs
[1347] MassARRAY Homogenous MassEXTEND.TM. (hME) Assay
[1348] The Homogenous MassEXTEND.TM. (hME) Assay uses a beadless,
label-free primer extension chemistry for genotyping. Each of the
primer products has a unique molecular weight that allows the
associated genotype to be precisely identified using mass
spectrometry.
Template Amplification
[1349] The isolated maternal DNA is amplified (2.5 ng) in a 5 .mu.l
volume using a 384-microtiter plate format. Any number of SNPs can
be amplified, either in a single reaction or in multiple reactions.
Representative primers that are used to amplify SNP TSC0271628
(A/G), which is located on chromosome 21, are provided below:
TABLE-US-00246 Upstream Primer: 5' AGGAAATTGTGAAGTA 3' (SEQ ID NO:
643) Downstream Primer: 5' TAACTCACTCACTATC 3' (SEQ ID NO: 644)
[1350] The primers can be longer or shorter in nucleotide sequence.
PCR conditions recommended by the makers of MassARRAY Homogenous
MassEXTEND Assay are followed. Representative PCR conditions are
provided below: [1351] (1) 95.degree. C. for 15 minutes and 15
seconds; [1352] (2) 95.degree. C. for 30 seconds; [1353] (4)
57.degree. C. for 30 seconds; [1354] (5) 72.degree. C. for 30
seconds; [1355] (6) Repeat steps 2-5 thirty two (32) times; [1356]
(7) 72.degree. C. for 5 minutes.
Dephosphorylation
[1357] Arctic shrimp alkaline phosphatase is added to the samples,
which are then incubated at 37.degree. C. for 20 minutes. This step
is done to dephosphorylate any remaining nucleotides, which
prevents their future incorporation and interference with MassARRAY
Homogenous MassEXTEND Assay. Samples are then incubated at
85.degree. C. to inactivate the heat-labile SAP.
hME Reaction
[1358] A MassEXTEND primer is designed to anneal close to the
polymorphic site, and is designed to identify both alleles of the
polymorphic site. For SNP TSC0271628, a representative MassEXTEND
primer is:
TABLE-US-00247 5' CTTTTTATGCCTTTCCACTCATCCA 3' (SEQ ID NO: 645)
[1359] The length of the MassEXTEND primer is designed according to
the instructions provided by the makers of the MassARRAY Homogenous
MassEXTEND Assay.
[1360] The MassEXTEND primer, DNA polymerase, and a cocktail
mixture of deoxynucleotides (dNTPs) and dideoxynucleotides (ddNTPs)
are added to the initial primer extension reaction. Allele-specific
primer products are generated that are generally one to four bases
longer than the original MassEXTEND primer.
[1361] A MassEXTEND primer is hybridized closely adjacent to the
polymorphic site following the conditions recommended by the makers
of the MassARRAY Homogenous MassEXTEND Assay. Nucleotide mixtures
are selected to maximize mass differences for all possible
MassEXTEND products. Appropriate dNTPS are incorporated until a
single ddNTP is incorporated, and the reaction is terminated. The
manufacturer's protocols are followed for all steps of the hME
assay.
[1362] Representative reaction products for SNP TSC0271628 are
provided below:
A Allele Before Primer Extension
TABLE-US-00248 [1363] (SEQ ID NO: 646) MassEXTEND primer: CT TTTT
ATGCCT T TCCACTCATCCA (SEQ ID NO: 647) Sample DNA:
GAAAAATACGGAAAGGTGAGTAGGTTTCC
[1364] The SNP site is identified in bold. After incubation with
DNA polymerase, ddATP, dCTP, dGTP, and dTTP, the following product
is generated:
[1365] A Allele after Primer Extension
TABLE-US-00249 (SEQ ID NO: 648) MassEXTEND primer: CT TTTT ATGCCT T
TCCACTCATCCAA* (SEQ ID NO: 649) Sample DNA:
GAAAAATACGGAAAGGTGAGTAGGTTTCC
[1366] ddATP is incorporated into the primer. Either labeled or
unlabeled ddNTPs can be used. The asterisk indicates ddATP that is
unlabeled. After the incorporation reaction, a 24-mer primer is
generated.
[1367] G Allele Before Primer Extension
TABLE-US-00250 (SEQ ID NO: 650) MassEXTEND primer: CT TTTT ATGCCT T
TCCACTCATCCA (SEQ ID NO: 651) Sample DNA:
GAAAAATACGGAAAGGTGAGTAGGTCTCC
[1368] The SNP site is identified in bold. After incubation with
DNA polymerase, ddATP, dCTP, dGTP, and dTTP, the following product
is generated:
[1369] G Allele after Primer Extension
TABLE-US-00251 (SEQ ID NO: 652) MassEXTEND primer: CT TTTT ATGCCT T
TCCACTCATCCAG A* (SEQ ID NO: 653) Sample DNA:
GAAAAATACGGAAAGGTGAGTAGGTCTCC
[1370] After the incorporation reaction, a 25-mer primer is
generated. The difference in molecular weight between the reaction
product for the A allele (24-mer) and the reaction product for the
G allele (25-mer) is used to genotype the locus of interest.
Sample Conditioning
[1371] SpectroCLEAN.TM. resin is added to the reaction to remove
extraneous salts that interfere with MALDI-TOF analysis.
Sample Transfer
[1372] 15 nl of sample is transferred from the 384-microtiter plate
and spotted onto the pad of the 384 SpectroCHIP.TM. microarray.
Sample Analysis
[1373] The SpectroCHIP.TM. is placed into the MALDI-TOF, which
measures the mass of the extension products. Once determined, the
genotype is called in real-time with SpectroTYPER.TM. RT software.
SNPs at which the maternal DNA are homozygous are identified, and
analyzed with the DNA that is isolated from the plasma.
Analysis of DNA Isolated from Maternal Plasma
[1374] After the maternal DNA is analyzed and homozygous SNPs are
identified, these SNPs are analyzed with the DNA isolated from the
plasma. A low copy number of fetal genomes typically exists in the
maternal plasma. To increase the copy number of the loci of
interest, which are the SNPs at which the maternal DNA is
homozygous, primers are designed to anneal at approximately 130
bases upstream and 130 bases downstream of each loci of interest.
This is done to reduce statistical sampling error that can occur
when working with a low number of genomes, which can influence the
ratio of one allele to another (see Example 11).
Design of Multiplex Primers
[1375] The primers are 12 bases in length. However, primers of any
length can be used including but not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers are designed to anneal to both the sense strand and
the antisense strand.
[1376] The maternal homozygous SNPs vary from sample to sample so
defined sequences are not provided here. Primers are designed to
anneal about 130 bases upstream and downstream of the maternal
homozygous SNPs. The primers are designed to terminate at the 3'
end in the dinucleotide "AA" to reduce the formation of
primer-dimers. However, the primers can be designed to end in any
of the four nucleotides and in any combination of the four
nucleotides.
Multiplex PCR
[1377] Regions upstream and downstream of the maternal homozygous
SNPs are amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
uses primers that anneal approximately 130 bases upstream and
downstream of each loci of interest. The primers are mixed together
and are used in a single reaction to amplify the template DNA. This
reaction is done to increase the number of copies of the loci of
interest, which eliminates error generated from a low number of
genomes.
[1378] For increased specificity, a "hot-start" PCR reaction is
used. PCR reactions are performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction is optimized for each locus of
interest. In this example, the 20 .mu.l of plasma template DNA is
used.
[1379] Two microliters of each forward and reverse primer, at
concentrations of 5 mM are pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix is used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.l of primer mix, and 25 .mu.l
of HotStar Taq. Twenty-five cycles of PCR are performed. The
following PCR conditions are used: [1380] (1) 95.degree. C. for 15
minutes; [1381] (2) 95.degree. C. for 30 second; [1382] (3)
4.degree. C. for 30 seconds; [1383] (4) 37.degree. C. for 30
seconds; [1384] (5) Repeat steps 2-4 twenty-four (24) times; [1385]
(6) 72.degree. C. for 10 minutes.
[1386] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
[1387] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
Purification of Fragment of Interest
[1388] The unused primers, and nucleotides are removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions are performed following the
manufacturer's instructions supplied with the columns. The DNA is
eluted in 100 .mu.l of sterile water. 5 .mu.l of each amplified
loci is mixed together
MassARRAY Homogenous MassEXTEND Assay
[1389] The pooled DNA is assayed with the hME assay as described
above. Each SNP is genotyped. SNPs located on chromosomes 13 and
21, wherein the maternal DNA is homozygous, and DNA isolated from
the plasma is heterozygous are quantitated.
Quantification
[1390] The intensity of each peak, wherein each peak corresponds to
a DNA fragment with a specific molecular weight, is quantitated. As
discussed above, the expected ratio of allele 1 to allele 2 is used
to determine the presence or absence of a chromosomal abnormality.
If the maternal genome is homozygous at SNP X (A/A), and the plasma
DNA is heterozygous at SNP X (A/G), then the G represents the
distinct fetal signal.
[1391] There will be some fragments that differ in molecular due to
the presence of the G nucleotide at SNP X in the fetal genome. The
intensity of the peak with the A nucleotide is quantitated and the
intensity of the peak that corresponds to fragments with the G
nucleotide is quantitated. The ratio of G:A depends on the
percentage of fetal DNA present in the maternal blood.
[1392] For example, if the sample contains 50% fetal DNA, then the
expected ratio is 0.33 (1 fetal G allele/(2 maternal A alleles+1
fetal A allele)). This ratio should be constant for all chromosomes
that are present in two copies. The ratio that is obtained for SNPs
on chromosome 13 should be the same as the ratio that is obtained
for chromosome 21.
[1393] However, if the fetal genome contains an additional copy of
chromosome 21, then the ratio for this chromosome will deviate from
the expected ratio. The expected ratio for a Trisomy condition with
50% fetal DNA in the maternal blood is 0.25. Thus, by analyzing
SNPs wherein the maternal genome is homozygous, and the DNA that is
isolated from the plasma is heterozygous, fetal chromosomal
abnormalities can be detected.
[1394] This example explained the use Sequenom's MassARRAY
Homogenous MassEXTEND (hME) assay, but it not intended to limit the
use of techniques that differentiate molecules based on molecular
weight.
Example 21
[1395] Fetal chromosomal abnormalities are determined by analyzing
SNPs wherein the maternal template DNA is homozygous and the
template DNA obtained from the plasma is heterozygous. Plasma that
is isolated from blood of a pregnant female contains both maternal
template DNA and fetal template DNA. Any number of SNP detection
methods can be used to analyze the maternal and plasma DNA. In this
example, SNPs are analyzed using Orchid's SNP-IT.TM. Assay.
However, other SNP detection methods based on primer extension may
also be used.
Collection of Blood Samples
[1396] In accordance with an IRB approved study, blood samples are
collected from pregnant women after informed consent is granted.
Blood is collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284) and 0.225 ml of 10% neutral buffered solution containing
formaldehyde (4% w/v), is added to each tube, and each tube gently
is inverted. The tubes are stored at 4.degree. C. until ready for
processing.
[1397] Any number of agents that impede cell lysis or stabilize
cell membranes can be added to the tubes including but not limited
to formaldehyde, and derivatives of formaldehyde, formalin,
glutaraldehyde, and derivatives of glutaraldehyde, crosslinkers,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP,
DTME, DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS or
compounds listed in Table XXIII. Any concentration of agent that
stabilizes cell membranes or impedes cell lysis can be added. In a
preferred embodiment, the agent that stabilizes cell membranes or
impedes cell lysis is added at a concentration that does not impede
or hinder subsequent reactions.
[1398] An agent that stabilizes cell membranes may be added to the
maternal blood sample to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, mexitil, dilantin, sodium hyaluronate, or polaxamer
188.
Isolation of Plasma and Maternal Cells
[1399] The blood is stored at 4.degree. C. until processing. The
tubes are spun at 1000 rpm for ten minutes in a centrifuge with
braking power set at zero. The tubes are spun a second time at 1000
rpm for ten minutes. The supernatant (the plasma) of each sample is
transferred to a new tube and spun at 3000 rpm for ten minutes with
the brake set at zero. The supernatant is transferred to a new tube
and stored at -80.degree. C. Approximately two milliliters of the
"bully coat," which contains maternal cells, is placed into a
separate tube and stored at -80.degree. C.
Isolation of DNA
[1400] DNA is isolated from the plasma sample using the Qiagen Midi
Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA is eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also is used to isolate DNA from the maternal cells
contained in the "buffy coat."
Identification of Homozygous Maternal SNPs
SNP-IT.TM. Assay
[1401] The SNP-IT Assay is based on a single base primer extension.
Prior to the SNP-IT Assay, a PCR product that includes the SNP of
interest is prepared, using one unmodified and one
phosphorothiolate modified primer. The PCR product is then rendered
single stranded using exonuclease, and the single stranded DNA is
annealed to a SNP-IT oligonucleotide immobilized on the surface of
a 96-well microtiter plate. Following hybridization, single base
extension occurs by the addition of DNA polymerase and the labeled
terminators. The incorporated base is detected using antibodies
specific to the label followed by colorimetric detection. Data
analysis can be done visually or with the use of the absorbance
plate reader.
Primer Design
[1402] For each locus of interest, the SNP-IT.TM. Assay requires
three primers. The primers are designed to produce an amplicon of
100-50 base pairs. The sequence of the SNP-IT primer, which is
designed to anneal immediately upstream of the SNP site, is the
best sequence available from between the upper and the lower
strands. The sequence of the SNP-IT primer is designed to minimize
hybridization to self and other sites in the amplicon. In addition,
the SNP-IT primer may contain modified bases to prevent
self-priming. The length of the primers is designed according to
the makers of the SNP-IT Assay.
[1403] Representative primers for the amplification and genotyping
of SNP TSC0069085, which is located on chromosome 21, are provided
below:
TABLE-US-00252 Upstream Primer: 5' ATCACACTGGGGATC 3' (SEQ ID NO:
654) Downstream Primer: 5' CTAAACCTATGACTC 3' (SEQ ID NO: 655)
SNP-IT primer 5' TTCACAGAGGATATCTTAATA 3' (SEQ ID NO: 656)
[1404] The upstream primer is unmodified and the downstream primer
is phosphorothiolate modified.
SNP-IT Plate Coating
[1405] SNP-IT primer is added to coat wells of empty 96-well
plates. This reaction typically incubates overnight. The
manufacturer's protocols and procedures are followed.
PCR
[1406] Template DNA (15 ng) is amplified either in a reaction
vessel including but not limited to an eppendorf tube or a well of
a microtiter plate. The manufacturer's protocols and procedures are
followed for the PCR reaction.
Exonuclease
[1407] PCR product is treated with exonuclease to degrade the
unmodified strand. The protected phosphorothiolate-labeled strand
is used in the SNP-IT Assay. The manufacturer's protocols and
procedures are followed for the exonuclease reaction.
Annealing
[1408] Single stranded PCR product transferred to SNP-IT plate and
is allowed to form a hybrid with the SNP-IT primer. The annealing
reaction typically proceeds for one hour. The manufacturer's
protocols and procedures are followed for the annealing
reaction.
SNP-IT Reaction
[1409] The extension reagent, which contains DNA polymerase, two
terminating nucleotides labeled with either fluorescein or biotin
and two unlabeled terminators, is added to the SNP-IT well
containing the annealed template and primer complex. For SNP
TSC0069085, ddCTP is labeled with fluorescein and ddTTP is labeled
with biotin, and the unlabeled terminators are ddATP, and ddGTP.
The manufacturer's protocols and procedures are followed for the
extension reaction.
[1410] The SNP specific base is incorporated by single base
extension of the SNP-IT primer. Primers are washed manually or in a
plate washer to remove unincorporated material. The manufacturer's
protocols and procedures are followed for the washing reaction.
Detection
[1411] Anti-fluorescein labeled with alkaline phosphatase (AP) is
added to the plate and allowed to bind to any incorporated
fluorescein labeled terminator. The manufacturer's protocols and
procedures for the labeling reaction are followed.
[1412] The plates are washed, and then color development is
performed using pNPP as the detection substrate. The absorbance is
read at 405 nm to detect yellow colored pNPP substrate followed by
a washing step to remove pNPP detection reagents. The
manufacturer's protocols and procedures are followed for color
development and washing steps.
[1413] Streptavidin labeled with horse radish peroxidase (HRP) is
added to the plate and allowed to bind to any incorporated biotin
labeled terminator. The manufacturer's protocols and procedures are
followed for the labeling reaction.
[1414] Following washing, color development is performed using TMB
as the detection substrate. The absorbance is read at 620 nm to
detect blue colored TMB substrate.
Analysis
[1415] Absorbance is plotted to generate a scatter plot from which
genotype calls are made. SNPs at which the maternal DNA is
homozygous are identified, and analyzed with the DNA isolated from
the maternal plasma.
Analysis of DNA Isolated from Maternal Plasma
[1416] After the maternal DNA is analyzed and homozygous SNPs are
identified, these SNPs are analyzed with the DNA isolated from the
plasma. A low copy number of fetal genomes typically exists in the
maternal plasma. To increase the copy number of the loci of
interest, which are the SNPs at which the maternal DNA is
homozygous, primers are designed to anneal at approximately 130
bases upstream and 130 bases downstream of each loci of interest.
This is done to reduce statistical sampling error that can occur
when working with a low number of genomes, which can influence the
ratio of one allele to another (see Example 11).
Design of Multiplex Primers
[1417] The primers are 12 bases in length. However, primers of any
length can be used including but not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers are designed to anneal to both the sense strand and
the antisense strand.
[1418] The maternal homozygous SNPs vary from sample to sample so
defined sequences are not provided here. Primers are designed to
anneal about 130 bases upstream and downstream of the maternal
homozygous SNPs. The primers are designed to terminate at the 3'
end in the dinucleotide "AA" to reduce the formation of
primer-dimers. However, the primers can be designed to end in any
of the four nucleotides and in any combination of the four
nucleotides.
Multiplex PCR
[1419] Regions upstream and downstream of the maternal homozygous
SNPs are amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
uses primers that anneal approximately 130 bases upstream and
downstream of each loci of interest. The primers are mixed together
and are used in a single reaction to amplify the template DNA. This
reaction is done to increase the number of copies of the loci of
interest, which eliminates error generated from a low number of
genomes.
[1420] For increased specificity, a "hot-start" PCR reaction is
used. PCR reactions are performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction is optimized for each locus of
interest. In this example, the 20 .mu.l of plasma template DNA is
used.
[1421] Two microliters of each forward and reverse primer, at
concentrations of 5 mM are pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix is used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.l of primer mix, and 25 .mu.l
of HotStar Taq. Twenty-five cycles of PCR are performed. The
following PCR conditions are used: [1422] (1) 95.degree. C. for 15
minutes; [1423] (2) 95.degree. C. for 30 second; [1424] (3)
4.degree. C. for 30 seconds; [1425] (4) 37.degree. C. for 30
seconds; [1426] (5) Repeat steps 2-4 twenty-four (24) times; [1427]
(6) 72.degree. C. for 10 minutes.
[1428] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
[1429] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
Purification of Fragment of Interest
[1430] The unused primers, and nucleotides are removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions are performed following the
manufacturer's instructions supplied with the columns. The DNA is
eluted in 100 .mu.l of sterile water. 5 .mu.l of each amplified
loci is mixed together
SNP-IT Assay
[1431] The pooled DNA is assayed with the SNP-IT assay as described
above. Each SNP is genotyped. SNPs located on chromosomes 13 and
21, wherein the maternal DNA is homozygous, and DNA isolated from
the plasma is heterozygous are quantitated.
Quantification
[1432] The fluorescence intensity of each allele is quantitated. As
discussed above, the expected ratio of allele 1 to allele 2 is used
to determine the presence or absence of a chromosomal abnormality.
If the maternal genome is homozygous at SNP X (A/A), and the plasma
DNA is heterozygous at SNP X (A/G), then the G represents the
distinct fetal signal.
[1433] The intensity of the allele with the A nucleotide is
quantitated and the intensity of the allele with the G nucleotide
is quantitated. The ratio of G:A depends on the percentage of fetal
DNA present in the maternal blood.
[1434] For example, if the sample contains 50% fetal DNA, then the
expected ratio is 0.33 (1 fetal G allele/(2 maternal A alleles+1
fetal A allele)). This ratio should be constant for all chromosomes
that are present in two copies. The ratio that is obtained for SNPs
on chromosome 13 should be the same as the ratio that is obtained
for chromosome 21.
[1435] However, if the fetal genome contains an additional copy of
chromosome 21, then the ratio for this chromosome will deviate from
the expected ratio. The expected ratio for a Trisomy condition with
50% fetal DNA in the maternal blood is 0.25. Thus, by analyzing
SNPs wherein the maternal genome is homozygous, and the DNA that is
isolated from the plasma is heterozygous, fetal chromosomal
abnormalities can be detected.
[1436] In this example, the terminator nucleotides are labeled with
different chemical moieties. However, using the methods described
in this application (see Example 6), the SNP-IT assay could be
modified to allow detection of both alleles with a single labeled
terminator.
[1437] This example explained the use of Orchid's SNP-IT assay, but
it not intended to limit the use of other techniques that rely on
primer extension. Orchid's SNPstream 25K, as well accompanying
software including but not limited to GetGenos.TM., QCreview.TM.,
and ValidGenos.TM., can also be used to detect the presence of
chromosomal abnormalities in the maternal blood. Additional
information about these products can be found at:
[1438]
http://www.orchidbio.com/products/lsg/products/snpstream.asp.
Example 22
[1439] Fetal chromosomal abnormalities are determined by analyzing
SNPs wherein the maternal template DNA is homozygous and the
template DNA obtained from the plasma is heterozygous. Plasma that
is isolated from blood of a pregnant female contains both maternal
template DNA and fetal template DNA. Any number of SNP detection
methods can be used to analyze the maternal and plasma DNA. In this
example, SNPs are analyzed using the TaqMan.RTM. assay. However,
other methods that rely on fluorogenic 5' nuclease assay can be
used.
Collection of Blood Samples
[1440] In accordance with an IRB approved study, blood samples are
collected from pregnant women after informed consent is granted.
Blood is collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284) and 0.225 ml of 10% neutral buffered solution containing
formaldehyde (4% w/v), is added to each tube, and each tube gently
is inverted. The tubes are stored at 4.degree. C. until ready for
processing.
[1441] Any number of agents that impede cell lysis or stabilize
cell membranes can be added to the tubes including but not limited
to formaldehyde, and derivatives of formaldehyde, formalin,
glutaraldehyde, and derivatives of glutaraldehyde, crosslinkers,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP, DTME,
DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS or compounds
listed in Table XXIII. Any concentration of agent that stabilizes
cell membranes or impedes cell lysis can be added. In a preferred
embodiment, the agent that stabilizes cell membranes or impedes
cell lysis is added at a concentration that does not impede or
hinder subsequent reactions.
[1442] An agent that stabilizes cell membranes may be added to the
maternal blood sample to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, mexitil, dilantin, sodium hyaluronate, or polaxamer
188.
Isolation of Plasma and Maternal Cells
[1443] The blood is stored at 4.degree. C. until processing. The
tubes are spun at 1000 rpm for ten minutes in a centrifuge with
braking power set at zero. The tubes are spun a second time at 1000
rpm for ten minutes. The supernatant (the plasma) of each sample is
transferred to a new tube and spun at 3000 rpm for ten minutes with
the brake set at zero. The supernatant is transferred to a new tube
and stored at -80.degree. C. Approximately two milliliters of the
"buffy coat," which contains maternal cells, is placed into a
separate tube and stored at -80.degree. C.
Isolation of DNA
[1444] DNA is isolated from the plasma sample using the Qiagen Midi
Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA is eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also is used to isolate DNA from the maternal cells
contained in the "buffy coat."
Identification of Homozygous Maternal SNPs
TaqMan Assay
[1445] PE Biosystems has two instruments in its Sequence Detection
Systems product line, the ABI Prism.RTM. 7700 Sequence Detection
System and the GeneAmp.RTM. 5700 Sequence Detection System. These
real-time systems allegedly are capable of detecting PCR products
as they accumulate during PCR and so enable the quantitation of DNA
in the sample.
[1446] One chemistry available for use on the ABI PRISM.RTM. 7700
and GeneAmp.RTM. 5700 detection systems is the fluorogenic 5'
nuclease assay, or the TaqMan.RTM. assay, which uses a fluorogenic
probe to enable the detection of a specific PCR product as it
accumulates during PCR. PE Biosystems' patented fluorogenic probe
design that incorporates the reporter due on the 5' end and the
quencher on the 3' end has assisted with the design of TaqMan
probes.
[1447] The basis for PCR quantitation in the ABI 7700 instrument is
to continuously measure PCR product accumulation using a
dual-labeled fluorogenic oligonucleotide probe called a TaqMan.RTM.
probe, which is composed of a short (20-25 bases)
oligodeoxynucleotide that is labeled with two different fluorescent
dyes. On the 5' terminus is a reporter dye and on the 3' terminus
is a quenching dye. This oligonucleotide probe sequence is
homologous to an internal sequence present in the PCR amplicon.
When the probe is intact, energy transfer occurs between the two
fluorophores and emission from the reporter is quenched by the
quencher (Livak et al., PCR Methods and Applications, 4:357-362,
1995a; U.S. Pat. No. 5,538,848; U.S. Pat. No. 5,723,591).
[1448] During the extension phase of PCR, the probe is cleaved by
5' nuclease activity of Taq polymerase thereby releasing the
reporter from the oligonucleotide-quencher and producing an
increase in reporter emission intensity. The ABI Prism 7700 uses
fiber optic systems which connect to each well in a 96-well PCR
tray format. The laser light source excites each well and a CCD
camera measures the fluorescence spectrum and intensity from each
well to generate real-time data during PCR amplification. The ABI
7700 Prism software examines the fluorescence intensity of reporter
and quencher dyes and calculates the increase in normalized
reporter emission intensity over the course of the amplification.
The results are then plotted versus time, represented by cycle
number, to produce a continuous measure of PCR amplification. To
provide precise quantification of initial target in each PCR
reaction, the amplification plot is examined at a point during the
early log phase of product accumulation. This is accomplished by
assigning a fluorescence threshold above background and determining
the time point at which each sample's amplification plot reaches
the threshold (defined as the threshold cycle number or CT).
Differences in threshold cycle number are used to quantify the
relative amount PCR target contained within each tube as described
previously.
[1449] For SNP analysis, a TaqMan probe can be designed for each
allele of the SNP. The reporter emission is used to determine the
presence or absence of each allele at the SNP. For example, for a
SNP that can either be adenine or guanine, a TaqMan probe will be
designed with a complementary nucleotide to the adenine and a
separate TaqMan probe will be designed with a complementary
nucleotide to the guanine. The two TaqMan probes can be used in
separate reaction vessels, which allows the amount of the adenine
allele and the amount of the guanine allele to be calculated.
Primer and Probe Design
[1450] Primer and probes can be designed using the Primer
Express.RTM. software. The probe is designed first, and then the
primers are designed as close as possible to the probe without
overlapping it. Amplicons of 50-150 base pairs are strongly
recommended.
[1451] The primer and probes should be designed following the
manufacturer's recommendations. For both the primer and the probes,
the G/C is in the range of 20-80%. The primer and probes are
designed to avoid runs of an identical nucleotide. This is
especially true for guanine, where runs of four or more Gs should
be avoided.
[1452] For the probe, the TM is about 68-70.degree. C., and is
designed so that there is no guanine on the 5' end. Also, the probe
is designed so that there are more C than G bases.
[1453] For the primers, the TM is about 58-60.degree. C., and the
primers are designed so that the five nucleotides at the 3' end
have no more than 2 G and/or C bases.
[1454] For example, representative primers and probes for SNP
TSC0271628 (A/G), which is located on chromosome 21, are provided
below:
TABLE-US-00253 Forward Primer (T.sub.M of 60.degree. C.) (SEQ ID
NO: 657) 5' AGTCTTGTAATACGACAGTCTT 3' Reverse Primer (T.sub.M of
58.degree. C.) (SEQ ID NO: 658) 5' CCATATCAATCAGTACTCTTG 3' TaqMan
Probe A allele (T.sub.M of 68.degree. C.; bold indicates variable
nucleotide at SNP) (SEQ ID NO: 659) 5' CCTTTCCACTCATCCAAAGGTTG 3'
TaqMan Probe G allele (T.sub.M of 70.degree. C.; bold indicates
variable nucleotide at SNP) (SEQ ID NO: 660) 5'
CCTTTCCACTCATCCAGAGGTTG 3'
[1455] Information regarding the sequence surrounding SNPs is found
at: http://www.snp.cshl.org. By independently varying forward and
reverse primer concentrations, the concentrations that provide
optimal assay conditions can be identified. Primer concentration
ranges of 50 nM-900 nM are tested.
[1456] If the maternal DNA is homozygous for one allele, for
example, adenine, then in the sample that contains the TaqMan probe
specific for the guanine nucleotide, the reporter is not separated
from the quencher because the TaqMan probe does not anneal to the
template DNA. However, if the maternal DNA is homozygous, then the
reporter will be separated from the quencher in both samples
containing the TaqMan probe specific for the guanine allele and
samples containing the TaqMan probe specific for the adenine
allele.
Reagent Solution
[1457] The polymerase recommended for the TaqMan Assay is AmpliTaq
Gold DNA polymerase. It is thought that the use of AmpliTaq Gold
DNA polymerase reduces the amount of non-specific product
formation. The incorporation of AmpErase.RTM. Uracil n-glycosylase
(UNG) and dUTP provide protection against PCR carryover
contamination. For PCR reactions, the TaqMan Universal PCR Master
mix, which is a reagent designed to provide optimal performance for
TaqMan assays, is recommended by the manufacturer.
[1458] The TaqMan reaction buffer contains 5.5 mM MgCl2, 200 nM
each of dATP, dCTP, dGTP, 400 nM dUTP, 0.5 U of uracyl DNA
glycosylase, and 1.25 U of AmpliTaq gold.
Thermal Cycling Parameters
[1459] PCR amplification and detection for all primer-probe
combinations are performed with the ABI 7700 Sequence Detection
System. The recommended cycling parameters for the TaqMan assay are
provided below:
[1460] 1) 50.degree. C. for 2 min;
[1461] 2) 95.degree. C. for 10 min;
[1462] 3) 95.degree. C. for 15 sec;
[1463] 4) 60.degree. C. for 1 min;
[1464] 5) Repeat steps 3-4 for 40 cycles.
TaqMan Quantitation
[1465] External standards are generated from know quantities of DNA
containing an adenine nucleotide at SNP TSC0271628 and a guanine
nucleotide at SNP TSC0271628, spanning 6 orders of magnitude (from
5.times.10.sup.0 to 5.times.10.sup.6 copies). The detection
threshold is set at 10 times the standard deviation of the mean
baseline emission calculated for PCR cycles 3 to 15 (Shifts et al.,
Applied and Environmental Microbiology, Vol. 67, No. 6, 2781-2789,
2001). Standard curves relating the threshold cycle to DNA
concentrations are generated with the ABI Prism 7700 software
(available from Perkin Elmer).
[1466] SNPs at which the maternal DNA is homozygous are identified,
and are analyzed with the DNA isolated from the plasma sample.
Analysis of DNA Isolated from Maternal Plasma
[1467] After the maternal DNA is analyzed and homozygous SNPs are
identified, these SNPs are analyzed with the DNA isolated from the
plasma. A low copy number of fetal genomes typically exists in the
maternal plasma. To increase the copy number of the loci of
interest, which are the SNPs at which the maternal DNA is
homozygous, primers are designed to anneal at approximately 130
bases upstream and 130 bases downstream of each loci of interest.
This is done to reduce statistical sampling error that can occur
when working with a low number of genomes, which can influence the
ratio of one allele to another (see Example 11).
Design of Multiplex Primers
[1468] The primers are 12 bases in length. However, primers of any
length can be used including but not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers are designed to anneal to both the sense strand and
the antisense strand.
[1469] The maternal homozygous SNPs vary from sample to sample so
defined sequences are not provided here. Primers are designed to
anneal about 130 bases upstream and downstream of the maternal
homozygous SNPs. The primers are designed to terminate at the 3'
end in the dinucleotide "AA" to reduce the formation of
primer-dimers. However, the primers can be designed to end in any
of the four nucleotides and in any combination of the four
nucleotides.
Multiplex PCR
[1470] Regions upstream and downstream of the maternal homozygous
SNPs are amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
uses primers that anneal approximately 130 bases upstream and
downstream of each loci of interest. The primers are mixed together
and are used in a single reaction to amplify the template DNA. This
reaction is done to increase the number of copies of the loci of
interest, which eliminates error generated from a low number of
genomes.
[1471] For increased specificity, a "hot-start" PCR reaction is
used. PCR reactions are performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction is optimized for each locus of
interest. In this example, the 20 .mu.l of plasma template DNA is
used.
[1472] Two microliters of each forward and reverse primer, at
concentrations of 5 mM are pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix is used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.l of primer mix, and 25 .mu.l
of HotStar Taq. Twenty-five cycles of PCR are performed. The
following PCR conditions are used: [1473] (1) 95.degree. C. for 15
minutes; [1474] (2) 95.degree. C. for 30 second; [1475] (3)
4.degree. C. for 30 seconds; [1476] (4) 37.degree. C. for 30
seconds; [1477] (5) Repeat steps 2-4 twenty-four (24) times; [1478]
(6) 72.degree. C. for 10 minutes.
[1479] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
[1480] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PCR.
[1481] It is important to ensure that the region amplified contains
annealing sequences for the oligonucleotide probes in the
BeadArray. Upon purchase of the BeadArray service, each SNP and the
primers used to analyze each SNP are identified. With this
knowledge, the multiplex primers are designed to encompass
annealing regions for the primers in the BeadArray.
Purification of Fragment of Interest
[1482] The unused primers, and nucleotides are removed from the
reaction by using Qiagen MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions are performed following the
manufacturer's instructions supplied with the columns. The DNA is
eluted in 100 .mu.l of sterile water.
TaqMan Assay
[1483] The amplified DNA is assayed with the TaqMan assay as
described above. Each SNP is genotyped. SNPs located on chromosomes
13 and 21, wherein the maternal DNA is homozygous, and DNA isolated
from the plasma is heterozygous are quantitated.
Quantification
[1484] The fluorescent intensity of the TaqMan allele specific
probe is quantitated. As discussed above, the expected ratio of
allele 1 to allele 2 is used to determine the presence or absence
of a chromosomal abnormality. If the maternal genome is homozygous
at SNP X (A/A), and the plasma DNA is heterozygous at SNP X (A/G),
then the G represents the distinct fetal signal.
[1485] The fluorescent intensity of the allele with the A
nucleotide is quantitated and the intensity of the allele with the
G nucleotide is quantitated. The ratio of G:A depends on the
percentage of fetal DNA present in the maternal blood.
[1486] For example, if the sample contains 50% fetal DNA, then the
expected ratio is 0.33 (1 fetal G allele/(2 maternal A alleles+1
fetal A allele)). This ratio should be constant for all chromosomes
that are present in two copies. The ratio that is obtained for SNPs
on chromosome 13 should be the same as the ratio that is obtained
for chromosome 21.
[1487] However, if the fetal genome contains an additional copy of
chromosome 21, then the ratio for this chromosome will deviate from
the expected ratio. The expected ratio for a Trisomy condition with
50% fetal DNA in the maternal blood is 0.25. Thus, by analyzing
SNPs wherein the maternal genome is homozygous, and the DNA that is
isolated from the plasma is heterozygous, fetal chromosomal
abnormalities can be detected.
[1488] This example explained the use of the TaqMan assay, but it
not intended to limit the use of other techniques that employ 5'
nuclease activity. For example, the SYBR.RTM. Green I double
stranded dye can also be used with the ABI PRISM 7700 Sequence
Detection System and the GeneAmp.RTM. 5700 Sequence Detection
system for determining the sequence of maternal and fetal DNA. The
SYBR.RTM. Green I double stranded dye assay may be used to detect
the presence of fetal chromosomal abnormalities in the maternal
blood.
[1489] SYBR.RTM. Green I double stranded dye is a highly specific
double-stranded DNA binding dye that allows the detection of
product accumulation during PCR. However, the SYBR.RTM. Green I
double stranded dye assay detects all double stranded DNA including
non-specific reaction products. The advantage of the SYBR.RTM.
Green I double stranded dye assay is that it does not require a
probe.
[1490] The same primers design parameters are recommended for both
the TaqMan Assay and the SYBR.RTM. Green I double stranded dye
assay (see Primer Design section in Example 21). The primer
optimization parameters recommended for the TaqMan assay should
also be followed for the SYBR.RTM. Green I double stranded dye
assay. In addition, no template controls should also be run with
the various concentrations of primers.
[1491] Furthermore, Applied Biosystems sell other products that may
be used to determine the sequence of maternal and fetal DNA
including but not limited to Assays-on-Demand.TM. SNP genotyping
products, and Assays-by-Design.sup.SM Service SNP genotyping
products.
[1492] Having now fully described the invention, it will be
understood by those of skill in the art that the invention can be
performed with a wide and equivalent range of conditions,
parameters, and the like, without affecting the spirit or scope of
the invention or any embodiment thereof.
[1493] All documents, e.g., scientific publications, patents and
patent publications recited herein are hereby incorporated by
reference in their entirety to the same extent as if each
individual document was specifically and individually indicated to
be incorporated by reference in its entirety. Where the document
cited only provides the first page of the document, the entire
document is intended, including the remaining pages of the
document.
Example 23
[1494] Fetal chromosomal abnormalities are determined by analyzing
SNPs wherein the maternal template DNA is homozygous and the
template DNA obtained from the plasma is heterozygous. Plasma that
is, isolated from blood of a pregnant female contains both maternal
template DNA and fetal template DNA. Any number of SNP detection
methods can be used to analyze the maternal and plasma DNA. In this
example, SNPs are analyzed by Third Wave Technologies' Invader.TM.
Assay for Nucleic Acid Detection. However, other techniques that
exploit and quantitate biological structures formed in the presence
of the correct sequence can be used.
Collection of Blood Samples
[1495] In accordance with an IRB approved study, blood samples are
collected from pregnant women after informed consent is granted.
Blood is collected into 9 ml EDTA Vacuette tubes (catalog number
NC9897284) and 0.225 ml of 10% neutral buffered solution containing
formaldehyde (4% w/v), is added to each tube, and each tube gently
is inverted. The tubes are stored at 4.degree. C. until ready for
processing.
[1496] Any number of agents that impede cell lysis or stabilize
cell membranes can be added to the tubes including but not limited
to formaldehyde, and derivatives of formaldehyde, formalin,
glutaraldehyde, and derivatives of glutaraldehyde, crosslinkers,
primary amine reactive crosslinkers, sulfhydryl reactive
crosslinkers, sulfhydryl addition or disulfide reduction,
carbohydrate reactive crosslinkers, carboxyl reactive crosslinkers,
photoreactive crosslinkers, cleavable crosslinkers, AEDP, APG,
BASED, BM(PEO).sub.3, BM(PEO).sub.4, BMB, BMDB, BMH, BMOE, BS3,
BSOCOES, DFDNB, DMA, DMP, DMS, DPDPB, DSG, DSP, DSS, DST, DTBP,
DTME, DTSSP, EGS, HBVS, sulfo-BSOCOES, Sulfo-DST, Sulfo-EGS or
compounds listed in Table XXIII. Any concentration of agent that
stabilizes cell membranes or impedes cell lysis can be added. In a
preferred embodiment, the agent that stabilizes cell membranes or
impedes cell lysis is added at a concentration that does not impede
or hinder subsequent reactions.
[1497] An agent that stabilizes cell membranes may be added to the
maternal blood samples to reduce maternal cell lysis including but
not limited to aldehydes, urea formaldehyde, phenol formaldehyde,
DMAE (dimethylaminoethanol), cholesterol, cholesterol derivatives,
high concentrations of magnesium, vitamin E, and vitamin E
derivatives, calcium, calcium gluconate, taurine, niacin,
hydroxylamine derivatives, bimoclomol, sucrose, astaxanthin,
glucose, amitriptyline, isomer A hopane tetral phenylacetate,
isomer B hopane tetral phenylacetate, citicoline, inositol, vitamin
B, vitamin B complex, cholesterol hemisuccinate, sorbitol, calcium,
coenzyme Q, ubiquinone, vitamin K, vitamin K complex, menaquinone,
zonegran, zinc, ginkgo biloba extract, diphenylhydantoin,
perftoran, polyvinylpyrrolidone, phosphatidylserine, tegretol,
PABA, disodium cromglycate, nedocromil sodium, phenyloin, zinc
citrate, mexitil, dilantin, sodium hyaluronate, or polaxamer
188.
Isolation of Plasma and Maternal Cells
[1498] The blood is stored at 4.degree. C. until processing. The
tubes are spun at 1000 rpm for ten minutes in a centrifuge with
braking power set at zero. The tubes are spun a second time at 1000
rpm for ten minutes. The supernatant (the plasma) of each sample is
transferred to a new tube and spun at 3000 rpm for ten minutes with
the brake set at zero. The supernatant is transferred to a new tube
and stored at -80.degree. C. Approximately two milliliters of the
"buffy coat," which contains maternal cells, is placed into a
separate tube and stored at -80.degree. C.
Isolation of DNA
[1499] DNA is isolated from the plasma sample using the Qiagen Midi
Kit for purification of DNA from blood cells, following the
manufacturer's instructions (QIAmp DNA Blood Midi Kit, Catalog
number 51183). DNA is eluted in 100 .mu.l of distilled water. The
Qiagen Midi Kit also is used to isolate DNA from the maternal cells
contained in the "buffy coat."
Identification of Homozygous Maternal SNPs
Third Wave Technologies Invader.TM. Assay
[1500] The Invader.TM. Assay, which was developed by Third Wave
Technologies (Madison, Wis.), is an isothermal, "PCR-free" approach
to the detection and quantitative analysis of DNA. The Invader
Assay produces and amplifies an unrelated signal only in the
presence of the correct target sequence. The Invader.TM. Assay
relies on a thermostable member of the structure-specific
archeabacterial flap endonuclease (FEN) family, which cleaves
nucleic acid molecules at specific sites based on structure rather
than sequence. When uses with structure forming probes for known
sequences, the enzymes cleave in a structure and target
sequence-specific manner. The nucleases used with Third Wave
Technologies' assays are referred to as "Cleavase.RTM."
enzymes.
[1501] The Invader.TM. Assay uses two target-specific
oligonucleotides to create the substrate complex recognized by
Cleavase Enzymes (L. DeFrancesco, The Scientist, 12(21):16, 1998).
The substrate complex is formed when an upstream Invader oligo and
a downstream signal probe hybridize in tandem to the nucleic acid.
The 3' end of the Invader oligo must overlap the hybridization site
of the signal probe by at least one base (Harrington et at, Genes
and Development, 8:1344-55, 1994). The 5' end of the signal probe
has additional unpaired bases to form a 5' flap. Cleavase enzymes
cleave the signal probe where it overlaps the Invader oligo,
releasing the 5' arm. Reaction mixtures contain excess signal probe
and are carried out near the melting temperature of the probe. Many
signal probes can be cleaved for each copy of the target without
temperature cycling.
[1502] The overlap between the Invader oligo and the signal probe
is important. A mismatch positioned at the site of the overlap will
block the cleavage by disrupting the overlap, which may allow
discrimination of SNPs and mutations.
[1503] The Invader assay utilizes two sequential cleavage steps.
The 5' arm of the signal probe released in the first reaction is
not detected directly. Rather, a secondary cleavage product is the
actual source of the signal, detected by fluorescence resonance
energy transfer (FRET). The primary cleavage product, which is the
5' arm released in the first reaction, is used as an Invader oligo
that hybridizes to a supplied FRET probe in the secondary reaction.
The FRET probe is labeled with two dyes: a donor fluorophore and a
quenching acceptor fluorophore. When the nuclease cleaves the
secondary probe, the two fluorophores are separated, quenching is
eliminated, and the enhanced fluorescence signal from the donor dye
is detected.
Probe Design
[1504] The following probes are designed to determine the sequence
at SNP TSC1172576 (T/A), which is located on chromosome 13:
TABLE-US-00254 Invader Oligo for T allele: 5' CATGCAGATATACCGCATAT
3' (SEQ ID NO: 661) Invader Oligo for A allele: 5'
CATGCAGATATACCGCATAA 3' (SEQ ID NO: 662)
[1505] Invader oligonucleotides are designed to be complementary to
an 18-22 base region immediately upstream of the signal probe, with
an additional one base at the 3' end that "invades" the region
hybridized to the signal probe by one base.
TABLE-US-00255 Signal probe for T allele: GGTAGCATCTCTCAGCACAAGAG
(SEQ ID NO: 663) Signal probe for A allele: GGTAGCATCACTCAGCACAAGAG
(SEQ ID NO: 664)
[1506] The signal probes are designed to contain a 3' region that
is complementary to the target sequence and a non-complementary 5'
arm (the underlined sequence above) that is used for detection. The
signal probes are labeled on the 5' end with 6-carboxyfluorescein
(TET), hexachloro-6-carboxyfluorescein (HEX), 6-carboxyfluorescein
(FAM). However, the 5' end can be labeled with any chemical moiety
including but not limited to radioisotope, fluorescent reporter
molecule, chemiluminescent reporter molecule, antibody, antibody
fragment, hapten, biotin, derivative of biotin, photobiotin,
iminobiotin, digoxigenin, avidin, enzyme, acridinium, sugar,
enzyme, apoenzyme, homopolymeric oligonucleotide, hormone,
ferromagnetic moiety, paramagnetic moiety, diamagnetic moiety,
phosphorescent moiety, luminescent moiety, electrochemiluminescent
moiety, chromatic moiety, moiety having a detectable electron spin
resonance, electrical capacitance, dielectric constant or
electrical conductivity, and combinations thereof
[1507] Both invader and signal probes are complementary to either
the sense or the antisense target DNA strand, depending on which
results in the formation of the least number of predictable
secondary structures.
PCR Amplification
[1508] A fragment of DNA that surrounds SNP TSC1172576 is amplified
by PCR. The sequence of the upstream and downstream primers is
provided below:
TABLE-US-00256 Upstream Primer: 5' TAGCAGAATCTCTCAT 3' (SEQ ID NO:
665) Downstream Primer: 5' AGAGTATCTCATTTGTT 3' (SEQ ID NO:
666)
[1509] Amplification reactions are performed in a final volume of
100 .mu.l of containing 2 .mu.l of genomic DNA, 35 pmol of each
primer, 50 .mu.m of each deoxynucleotide (Perkin-Elmer Applied
Biosystems, Inc., Foster City, Calif.), 1.times.PCR buffer (20 mM
Tris-Hcl, 50 mM KCl, 1.5 mM MgCl.sub.2, 0.05% Tween-20, 0.05%
NP40), 1 M betaine, 5% dimethylsulfoxide (DMSO), and 2.5 U of Taq
polymerase (Roche Boehringer Mannheim, Indianapolis, Ind.). PCR
cycling conditions consist of an initial denaturation step at
95.degree. C. for 5 min, 30 cycles of denaturation at 95.degree. C.
for 1 min, annealing at 68.degree. C. for 1 min, and extension at
72.degree. C. for 1 min, and a final extension at 72.degree. C. for
5 min.
Invader Reaction
[1510] One microliter of each PCR product is added to 0.5 pmol of
the appropriate Invader oligonucleotide, 10 ng human genomic DNA
(Promega Corp., Madison, Wis.) as the carrier, and
mopholinepropanesulfonic acid (MOPS) buffer (pH 8.0) at a final
concentration of 10 mM in a volume of 7 .mu.l. The mixtures are
denatured for 5 min at 95.degree. C. and then cooled to reaction
temperature of 60.degree. C. Invader reactions are initiated by the
addition of a mixture containing 30 ng of Cleavase VIII (Third Wave
Technologies, Inc., Madison, Wis.) 25 mM MgCl.sub.2, and 10 pmol of
the appropriate signal probe oligonucleotide in a volume of 3
.mu.l. Reaction mixtures are incubated for 60 min. The reactions
are terminated by the addition of 10 .mu.l of 95% formamide--10 mM
EDTA (pH 8.0)--0.05% crystal violet. Following termination, the
reactions are diluted 1:10 in reagent-grade water. Samples of 2
.mu.l are loaded and electrophoresed in a 24% denaturing
polyacrylamide gel (18 cm by 25.5 cm by mm) on an automated
fluorescence sequencing apparatus (model 377, PE-ABI). The data are
collected using filter set C and processed with GeneScan
software.
[1511] In addition, 5 .mu.l of each sample is electrophoresed in
20% (acrylamide to bisacrylamide, 19:1) denaturing polyacrylamide
gels at 20 W. Gel cassettes (20 cm by 20 cm by 0.5 mM) are scanned
with a fluorescent scanner (FMBIO-100; Hitachi Corp, San Bruno,
Calif.) by using a 585 nm filter for TET and FLEX labeled probes
and a 505 nm titter for FAM labeled probes.
[1512] With the Invader assay, it is also possible to perform a
second cleavage reaction where the released 5' arm of the signal
probe is hybridized to another probe, and fluorescence resonance
energy transfer (FRET) is used to detect the presence of a specific
nucleic acid. The manufacturer's protocols are followed when using
the FRET probe.
[1513] The genotype at each SNP is determined by analyzing the
fluorescence intensity of each allele-specific signal probe. For
example, for SNP TSC1172576, the presence of the T allele is
determined by analyzing the amount of released 5' signal probe
using the signal probe from the T allele signal probe (as described
above). Likewise, the presence of the A allele is determined by
analyzing the amount of released 5' signal probe from the A allele
signal probe. The reactions can be performed in a single reaction
vessel using two different chemical moieties, which can be analyzed
under distinct conditions, or the A and T allele reactions can be
performed in two different reaction vessels.
Analysis of DNA Isolated from Maternal Plasma
[1514] After the maternal DNA is analyzed and homozygous SNPs are
identified, these SNPs are analyzed with the DNA isolated from the
plasma. A low copy number of fetal genomes typically exist in the
maternal plasma. To increase the copy number of the loci of
interest, which are the SNPs at which the maternal DNA is
homozygous, primers are designed to anneal at approximately 130
bases upstream and 130 bases downstream of each loci of interest.
This is done to reduce statistical sampling error that can occur
when working with a low number of genomes, which can influence the
ratio of one allele to another (see Example 11).
Design of Multiplex Primers
[1515] The primers are 12 bases in length. However, primers of any
length can be used including but not limited to 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36-45, 46-55, 56-65,
66-75, 76-85, 86-95, 96-105, 106-115, 116-125, and greater than 125
bases. Primers are designed to anneal to both the sense strand and
the antisense strand.
[1516] The maternal homozygous SNPs vary from sample to sample so
defined sequences are not provided here. Primers are designed to
anneal about 130 bases upstream and downstream of the maternal
homozygous SNPs. The primers are designed to terminate at the 3'
end in the dinucleotide "AA" to reduce the formation of
primer-dimers. However, the primers can be designed to end in any
of the four nucleotides and in any combination of the four
nucleotides.
Multiplex PCR
[1517] Regions upstream and downstream of the maternal homozygous
SNPs are amplified from the template genomic DNA using the
polymerase chain reaction (PCR, U.S. Pat. Nos. 4,683,195 and
4,683,202, incorporated herein by reference). This PCR reaction
uses primers that anneal approximately 130 bases upstream and
downstream of each loci of interest. The primers are mixed together
and are used in a single reaction to amplify the template DNA. This
reaction is done to increase the number of copies of the loci of
interest, which eliminates error generated from a low number of
genomes.
[1518] For increased specificity, a "hot-start" PCR reaction is
used. PCR reactions are performed using the HotStarTaq Master Mix
Kit supplied by QIAGEN (catalog number 203443). The amount of
template DNA and primer per reaction is optimized for each locus of
interest. In this example, the 20 .mu.l of plasma template DNA is
used.
[1519] Two microliters of each forward and reverse primer, at
concentrations of 5 mM are pooled into a single microcentrifuge
tube and mixed. Four microliters of the primer mix is used in a
total PCR reaction volume of 50 .mu.l (20 .mu.l of template plasma
DNA, 1 .mu.l of sterile water, 4 .mu.l of primer mix, and 25 .mu.l
of HotStar Taq. Twenty-five cycles of PCR are performed. The
following PCR conditions are used: [1520] (1) 95.degree. C. for 15
minutes; [1521] (2) 95.degree. C. for 30 second; [1522] (3)
4.degree. C. for 30 seconds; [1523] (4) 37.degree. C. for 30
seconds; [1524] (5) Repeat steps 2-4 twenty-four (24) times; [1525]
(6) 72.degree. C. for 10 minutes.
[1526] The temperatures and times for denaturing, annealing, and
extension, are optimized by trying various settings and using the
parameters that yield the best results.
[1527] Other methods of genomic amplification can also be used to
increase the copy number of the loci of interest including but not
limited to primer extension preamplification (PEP) (Zhang et al.,
PNAS, 89:5847-51, 1992), degenerate oligonucleotide primed PCR
(DOP-PCR) (Telenius, et al., Genomics 13:718-25, 1992), strand
displacement amplification using DNA polymerase from bacteriophage
29, which undergoes rolling circle replication (Dean et al.,
Genomic Research 11:1095-99, 2001), multiple displacement
amplification (U.S. Pat. No. 6,124,120), REPLI-g.TM. Whole Genome
Amplification kits, and Tagged PER.
[1528] It is important to ensure that the region amplified contains
annealing sequences for the oligonucleotide probes in the
BeadArray. Upon purchase of the BeadArray service, each SNP and the
primers used to analyze each SNP are identified. With this
knowledge, the multiplex primers are designed to encompass
annealing regions for the primers in the BeadArray.
Purification of Fragment of Interest
[1529] The unused primers, and nucleotides are removed from the
reaction by using Qiagen. MinElute PCR purification kits (Qiagen,
Catalog Number 28004). The reactions are performed following the
manufacturer's instructions supplied with the columns. The DNA is
eluted in 100 .mu.l of sterile water.
Invader Assay
[1530] The amplified DNA is assayed with the Invader assay as
described above. Each SNP is genotyped. SNPs located on chromosomes
13 and 21, wherein the maternal DNA is homozygous, and DNA isolated
from the plasma is heterozygous are quantitated.
Quantification
[1531] The fluorescent intensity of the allele specific signal
probe is quantitated. As discussed above, the expected ratio of
allele 1 to allele 2 is used to determine the presence or absence
of a chromosomal abnormality. If the maternal genome is homozygous
at SNP X (A/A), and the plasma DNA is heterozygous at SNP X (A/G),
then the G represents the distinct fetal signal.
[1532] The fluorescent intensity of the allele with the A
nucleotide is quantitated and the intensity of the allele with the
G nucleotide is quantitated. The ratio of G:A depends on the
percentage of fetal DNA present in the maternal blood.
[1533] For example, if the sample contains 50% fetal DNA, then the
expected ratio is 0.33 (1 fetal G allele/(2 maternal A alleles+1
fetal A allele)). This ratio should be constant for all chromosomes
that are present in two copies. The ratio that is obtained for SNPs
on chromosome 13 should be the same as the ratio that is obtained
for chromosome 21.
[1534] However, if the fetal genome contains an additional copy of
chromosome 21, then the ratio for this chromosome will deviate from
the expected ratio. The expected ratio for a Trisomy condition with
50% fetal DNA in the maternal blood is 0.25. Thus, by analyzing
SNPs wherein the maternal genome is homozygous, and the DNA that is
isolated from the plasma is heterozygous, fetal chromosomal
abnormalities can be detected.
Sequence CWU 1 SEQUENCE LISTING <160> NUMBER OF SEQ ID
NOS: 711 <210> SEQ ID NO 1 <211> LENGTH: 15 <212>
TYPE: DNA <213> ORGANISM: Unknown <220> FEATURE:
<223> OTHER INFORMATION: Restriction site <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(15) <223>
OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE: 1
gggacnnnnn nnnnn 15 <210> SEQ ID NO 2 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Unknown <220>
FEATURE: <223> OTHER INFORMATION: Restriction site
<221> NAME/KEY: misc_feature <222> LOCATION: (1)...(14)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
2 nnnnnnnnnn nnnngtccc 19 <210> SEQ ID NO 3 <211>
LENGTH: 21 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 3 ggaaattcca tgatgcgtgg g 21 <210> SEQ
ID NO 4 <211> LENGTH: 23 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (19)...(21) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 4 ggaaattcca
tgatgcgtnn nac 23 <210> SEQ ID NO 5 <211> LENGTH: 21
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 5 ggaaattcca tgatgcgtac c 21 <210> SEQ
ID NO 6 <211> LENGTH: 25 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (22)...(23) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 6 ggaaattcca
tgatgcgtac cnngg 25 <210> SEQ ID NO 7 <211> LENGTH: 11
<212> TYPE: DNA <213> ORGANISM: Unknown <220>
FEATURE: <223> OTHER INFORMATION: Restriction site
<221> NAME/KEY: misc_feature <222> LOCATION: (4)...(8)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
7 cctnnnnnag g 11 <210> SEQ ID NO 8 <211> LENGTH: 25
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (20)...(23) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 8 ggaaattcca tgatgcgtan nnngg 25 <210>
SEQ ID NO 9 <211> LENGTH: 38 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 9
tagaatagca ctgaattcag gaatacaatc attgtcac 38 <210> SEQ ID NO
10 <211> LENGTH: 28 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 10 atcacgataa
acggccaaac tcaggtta 28 <210> SEQ ID NO 11 <211> LENGTH:
38 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 11 aagtttagat cagaattcgt gaaagcagaa gttgtctg
38 <210> SEQ ID NO 12 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 12 tctccaacta acggctcatc gagtaaag 28 <210> SEQ ID
NO 13 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 13 atgactagct
atgaattcgt tcaaggtaga aaatggaa 38 <210> SEQ ID NO 14
<211> LENGTH: 28 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 14 gagaattaga acggcccaaa
tcccactc 28 <210> SEQ ID NO 15 <211> LENGTH: 37
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 15 ttacaatgca tgaattcatc ttggtctctc aaagtgc
37 <210> SEQ ID NO 16 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 16 tggaccataa acggccaaaa actgtaag 28 <210> SEQ ID
NO 17 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 17 ataaccgtat
gcgaattcta taattttcct gataaagg 38 <210> SEQ ID NO 18
<211> LENGTH: 28 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 18 cttaaatcag gggactaggt
aaacttca 28 <210> SEQ ID NO 19 <211> LENGTH: 28
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 19 cttaaatcag acggctaggt aaacttca 28
<210> SEQ ID NO 20 <211> LENGTH: 28 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 20
tctccaacta gggactcatc gagtaaag 28 <210> SEQ ID NO 21
<211> LENGTH: 37 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 21 aacgccgggc gagaattcag
tttttcaact tgcaagg 37 <210> SEQ ID NO 22 <211> LENGTH:
28 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 22 ctacacatat ctgggacgtt ggccatcc 28
<210> SEQ ID NO 23 <211> LENGTH: 38 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 23
taccttttga tcgaattcaa ggccaaaaat attaagtt 38 <210> SEQ ID NO
24 <211> LENGTH: 28 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 24 tcgaacttta
acggccttag agtagaga 28 <210> SEQ ID NO 25 <211> LENGTH:
38 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 25 cgatttcgat aagaattcaa aagcagttct tagttcag
38 <210> SEQ ID NO 26 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 26 tgcgaatctt acggctgcat cacattca 28 <210> SEQ ID
NO 27 <211> LENGTH: 23 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (3)...(5) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 27 gtnnnacgca
tcatggaatt tcc 23 <210> SEQ ID NO 28 <211> LENGTH: 25
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (3)...(4) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 28 ccnnggtacg catcatggaa tttcc 25 <210>
SEQ ID NO 29 <211> LENGTH: 25 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (3)...(6) <223>
OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE: 29
ccnnnntacg catcatggaa tttcc 25 <210> SEQ ID NO 30 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 30 gggctagtct ccgaattcca cctatcctac caaatgtc
38 <210> SEQ ID NO 31 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 31 tagctgtagt tagggactgt tctgagcac 29 <210> SEQ ID
NO 32 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 32 cgaatgcaag
gcgaattcgt tagtaataac acagtgca 38 <210> SEQ ID NO 33
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 33 aagactggat ccgggaccat
gtagaatac 29 <210> SEQ ID NO 34 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 34 tctaaccatt gcgaattcag ggcaaggggg gtgagatc
38 <210> SEQ ID NO 35 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 35 tgacttggat ccgggacaac gactcatcc 29 <210> SEQ ID
NO 36 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 36 acccaggcgc
cagaattctt tagataaagc tgaaggga 38 <210> SEQ ID NO 37
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 37 gttacgggat ccgggactcc
atattgatc 29 <210> SEQ ID NO 38 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 38 cgttggcttg aggaattcga ccaaaagagc caagagaa
38 <210> SEQ ID NO 39 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 39 aaaaagggat ccgggacctt gactaggac 29 <210> SEQ ID
NO 40 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 40 acttgattcc
gtgaattcgt tatcaataaa tcttacat 38 <210> SEQ ID NO 41
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 41 caagttggat ccgggaccca
gggctaacc 29 <210> SEQ ID NO 42 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 42 gtgcaaaggc ctgaattccc aggcacaaag ctgttgaa
38 <210> SEQ ID NO 43 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 43 tgaagcgaac tagggactca ggtggactt 29 <210> SEQ ID
NO 44 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 44 gattccgtaa
acgaattcag ttcattatca tctttgtc 38 <210> SEQ ID NO 45
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 45 ccattgttaa gcgggacttc
tgctatttg 29 <210> SEQ ID NO 46 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<400> SEQUENCE: 46 cccaaaagtc cacctga 17 <210> SEQ ID
NO 47 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <400> SEQUENCE: 47 tcaggtggac ttttggg
17 <210> SEQ ID NO 48 <211> LENGTH: 18 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <400> SEQUENCE:
48 accctgcaaa tagcagaa 18 <210> SEQ ID NO 49 <211>
LENGTH: 18 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<400> SEQUENCE: 49 ttctgctatt tgcagggt 18 <210> SEQ ID
NO 50 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <400> SEQUENCE: 50 acccgcaaat agcagaa
17 <210> SEQ ID NO 51 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <400> SEQUENCE:
51 ttctgctatt tgcgggt 17 <210> SEQ ID NO 52 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 52 ttagatagca
gtaattt 17 <210> SEQ ID NO 53 <211> LENGTH: 23
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(13) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 53 ggaagccggg
aaggatctgt atc 23 <210> SEQ ID NO 54 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 54 gagaaagaga ggtaa 15 <210> SEQ
ID NO 55 <211> LENGTH: 19 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(8) <223> OTHER
INFORMATION: These nucleotides may be absent <221> NAME/KEY:
misc_feature <222> LOCATION: (10)...(19) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
55 aaagagaggt aacttttct 19 <210> SEQ ID NO 56 <211>
LENGTH: 18 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 56 aaagagaggt
aacttttc 18 <210> SEQ ID NO 57 <211> LENGTH: 23
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 11 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 57 ttttaaaaaa aaaaaatagg tca 23
<210> SEQ ID NO 58 <211> LENGTH: 22 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION:
(11)...(12) <223> OTHER INFORMATION: These nucleotides may be
absent <400> SEQUENCE: 58 aaaataggtc attgcttctt gc 22
<210> SEQ ID NO 59 <211> LENGTH: 16 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (5)...(6)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 59 gacaaagaag aaaagg 16 <210> SEQ ID NO
60 <211> LENGTH: 19 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(9) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
60 gacaaagaag aaaaggaaa 19 <210> SEQ ID NO 61 <211>
LENGTH: 25 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (11)...(15) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 61 aggaaaaaga
ctggtattac gctca 25 <210> SEQ ID NO 62 <211> LENGTH: 24
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (11)...(14) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 62 aaaagaatag
atagtcttcc ttta 24 <210> SEQ ID NO 63 <211> LENGTH: 21
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 11 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 63 agatagtctt cctttaactg a 21
<210> SEQ ID NO 64 <211> LENGTH: 19 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(9)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 64 tccttacaaa cagatatga 19 <210> SEQ ID
NO 65 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 65 accagaaggc
aatt 14 <210> SEQ ID NO 66 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(7)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 66 atcagagttg cgatgga 17 <210> SEQ ID
NO 67 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 67 cgagcacagg
taagtt 16 <210> SEQ ID NO 68 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 68 cactctgcac ctcga
15 <210> SEQ ID NO 69 <211> LENGTH: 14 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: 4
<223> OTHER INFORMATION: This nucleotide may be absent
<400> SEQUENCE: 69 gatatgtcgc gaac 14 <210> SEQ ID NO
70 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
70 aaagactctg tattgtt 17 <210> SEQ ID NO 71 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 71 gacaagagag gcagg
15 <210> SEQ ID NO 72 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(7)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 72 catgaaccag gcatgga 17 <210> SEQ ID
NO 73 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 73 gaaccaggca
tggacc 16 <210> SEQ ID NO 74 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 74 aatccaagta
tgttctct 18 <210> SEQ ID NO 75 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 75 gctcctgttg aacatc 16 <210>
SEQ ID NO 76 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE: 76
aaactttcat ttgatg 16 <210> SEQ ID NO 77 <211> LENGTH:
19 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 77 aaactttcat
ttgatgaag 19 <210> SEQ ID NO 78 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 78 ctacaggcca ttgc 14 <210> SEQ
ID NO 79 <211> LENGTH: 21 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 11 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE: 79
taaattaggg ggactacagg c 21 <210> SEQ ID NO 80 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 80 ttattgcaag tggac 15 <210> SEQ
ID NO 81 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 81 tacgggctta
ctaat 15 <210> SEQ ID NO 82 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 82 agtattacac taagac 16 <210>
SEQ ID NO 83 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 83 attacactaa gacgata 17 <210> SEQ ID NO 84
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 84 ctaagacgat
atgc 14 <210> SEQ ID NO 85 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(7)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 85 tgctctatga aaggctg 17 <210> SEQ ID
NO 86 <211> LENGTH: 25 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(15) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
86 atgagagcac ttgtggccca actaa 25 <210> SEQ ID NO 87
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 87 gacttacagc
aggtac 16 <210> SEQ ID NO 88 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 88 aaaaagacgt
tgcgaga 17 <210> SEQ ID NO 89 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 89 gttggaagtg
tgaaagcat 19 <210> SEQ ID NO 90 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 90 aaagcattga tggaat 16 <210>
SEQ ID NO 91 <211> LENGTH: 18 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(8) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 91 ttagaagtta aaaaggta 18 <210> SEQ ID NO 92
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 92 accctcaaaa
gcgtat 16 <210> SEQ ID NO 93 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 93 gccttatgga atttg 15 <210> SEQ
ID NO 94 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 94 gctgtagatg
gtgc 14 <210> SEQ ID NO 95 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (4)...(19)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 95 gttggcactc ttacttaccg gagccagac 29
<210> SEQ ID NO 96 <211> LENGTH: 19 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(9)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 96 cttacttacc ggagccaga 19 <210> SEQ ID
NO 97 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 97 acttaccgga
gccag 15 <210> SEQ ID NO 98 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 98 agccagacaa acact
15 <210> SEQ ID NO 99 <211> LENGTH: 18 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (5)...(8)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 99 agccagacaa acacttta 18 <210> SEQ ID
NO 100 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
100 acaaacactt tagccat 17 <210> SEQ ID NO 101 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 101 ttagccatta ttgaaa 16 <210>
SEQ ID NO 102 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
102 ggaggtggga tatta 15 <210> SEQ ID NO 103 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 103 atattacgga atgtg 15 <210>
SEQ ID NO 104 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
104 ttacggaatg tgtcca 16 <210> SEQ ID NO 105 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 105 agagagaaca
actgt 15 <210> SEQ ID NO 106 <211> LENGTH: 34
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (11)...(24) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 106 tatttcaggc
aaatcctaag agagaacaac tgtc 34 <210> SEQ ID NO 107 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 107 aactgtctac
aaactt 16 <210> SEQ ID NO 108 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 108 ttattacaac
actta 15 <210> SEQ ID NO 109 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 109 cacttaaaat
ctcat 15 <210> SEQ ID NO 110 <211> LENGTH: 24
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(14) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 110 agtttgacaa
tagtcagtaa tgca 24 <210> SEQ ID NO 111 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 111 cacttatcag aaactt 16 <210>
SEQ ID NO 112 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
112 ttatcagaaa ctttt 15 <210> SEQ ID NO 113 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 113 tcagaaactt ttgaca 16 <210>
SEQ ID NO 114 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
114 agtcccaagg catct 15 <210> SEQ ID NO 115 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 115 aagcaaagtc tctat 15 <210>
SEQ ID NO 116 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 116 aagcaaagtc tctatgg 17 <210> SEQ ID NO 117
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
These nucleotides may be absent <400> SEQUENCE: 117
caaagtctct atggt 15 <210> SEQ ID NO 118 <211> LENGTH:
16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 118 gattatgttt
ttgaca 16 <210> SEQ ID NO 119 <211> LENGTH: 25
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(15) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 119 gacaccaatc
gacatgatga taata 25 <210> SEQ ID NO 120 <211> LENGTH:
16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 120 cgacatgatg
ataata 16 <210> SEQ ID NO 121 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 121 tcagacaatt
ttaatact 18 <210> SEQ ID NO 122 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 122 tatttgaata ctac 14 <210> SEQ
ID NO 123 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 123 aatactacag
tgtta 15 <210> SEQ ID NO 124 <211> LENGTH: 28
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(18) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 124 gtgttaccca
gctcctcttc atcaagag 28 <210> SEQ ID NO 125 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 125 agctcctctt catcaa 16 <210>
SEQ ID NO 126 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
126 tcatcaagag gaagc 15 <210> SEQ ID NO 127 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 127 aaagatagaa
gtttgga 17 <210> SEQ ID NO 128 <211> LENGTH: 21
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(11) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 128 aaagatagaa
gtttggagag a 21 <210> SEQ ID NO 129 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 129 gaacgcggaa ttggt 15 <210>
SEQ ID NO 130 <211> LENGTH: 19 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(9) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 130 cgcggaattg gtctaggca 19 <210> SEQ ID NO 131
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 131 cgcggaattg
gtcta 15 <210> SEQ ID NO 132 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 132 cagatctcca ccac 14 <210> SEQ
ID NO 133 <211> LENGTH: 19 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(9) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
133 gaagacagaa gttctgggt 19 <210> SEQ ID NO 134 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 134 gggtctacca ctgaa 15 <210>
SEQ ID NO 135 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
135 gtgacagatg agagaa 16 <210> SEQ ID NO 136 <211>
LENGTH: 19 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 136 catacacatt
caaacactt 19 <210> SEQ ID NO 137 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 137 acacattcaa
acacttaca 19 <210> SEQ ID NO 138 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 138 cattcaaaca ctta 14 <210> SEQ
ID NO 139 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 139 cattcaaaca
cttac 15 <210> SEQ ID NO 140 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 140 aacacttaca
atttcac 17 <210> SEQ ID NO 141 <211> LENGTH: 22
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(12) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 141 tacaatttca
ctaagtcgga aa 22 <210> SEQ ID NO 142 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 142 ttcactaagt
cggaaaat 18 <210> SEQ ID NO 143 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 143 aagtcggaaa
attcaaa 17 <210> SEQ ID NO 144 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 144 acatgttcta tgcct 15 <210>
SEQ ID NO 145 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 145 ttagaataca agagat 16 <210> SEQ ID NO 146
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 146 aatgatagtt
taaa 14 <210> SEQ ID NO 147 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 147 agtttaaata gtgtca 16 <210>
SEQ ID NO 148 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 148 ttaaatagtg tcagtag 17 <210> SEQ ID NO 149
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 149 tatggtaaaa
gaggt 15 <210> SEQ ID NO 150 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 150 ggtaaaagag gtcaaa 16 <210>
SEQ ID NO 151 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 151 aaaagaggtc aaatga 16 <210> SEQ ID NO 152
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 152 agtaagtttt
gcagtt 16 <210> SEQ ID NO 153 <211> LENGTH: 24
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(14) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 153 aagttttgca
gttatggtca atac 24 <210> SEQ ID NO 154 <211> LENGTH: 20
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(10) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 154 caatacccag
ccgacctagc 20 <210> SEQ ID NO 155 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 155 acaccaataa attat 15 <210>
SEQ ID NO 156 <211> LENGTH: 14 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 4 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
156 aaatattcag atga 14 <210> SEQ ID NO 157 <211>
LENGTH: 19 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 157 tcagatgagc
agttgaact 19 <210> SEQ ID NO 158 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 158 gatgagcagt tgaac 15 <210>
SEQ ID NO 159 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
159 tgggcaagac ccaaa 15 <210> SEQ ID NO 160 <211>
LENGTH: 19 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 160 cacataatag
aagatgaaa 19 <210> SEQ ID NO 161 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 161 ataatagaag
atgaaataa 19 <210> SEQ ID NO 162 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 162 atagaagatg aaataa 16 <210>
SEQ ID NO 163 <211> LENGTH: 20 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(10) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 163 ataaaacaaa gtgagcaaag 20 <210> SEQ ID NO 164
<211> LENGTH: 17 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
164 aaacaaagtg agcaaag 17 <210> SEQ ID NO 165 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 165 aaacaaagtg
agcaaa 16 <210> SEQ ID NO 166 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 166 caaagtgagc
aaagacaa 18 <210> SEQ ID NO 167 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 167 caaagacaat
caaggaa 17 <210> SEQ ID NO 168 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 168 caatcaagga
atcaaag 17 <210> SEQ ID NO 169 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 169 caaagtacaa
cttatc 16 <210> SEQ ID NO 170 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 170 actgagagca
ctgatg 16 <210> SEQ ID NO 171 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 171 actgatgata
aacacct 17 <210> SEQ ID NO 172 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 172 gataaacacc
tcaagtt 17 <210> SEQ ID NO 173 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 173 cacctcaagt
tccaac 16 <210> SEQ ID NO 174 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 174 tttggacagc aggaa 15 <210>
SEQ ID NO 175 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(5) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 175 tgtgtttctc catac 15 <210> SEQ ID NO 176
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 176 cggggagcca
atgg 14 <210> SEQ ID NO 177 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 177 tcagaaacaa atcgag 16 <210>
SEQ ID NO 178 <211> LENGTH: 19 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(9) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 178 attaatcaaa atgtaagcc 19 <210> SEQ ID NO 179
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 179 caagaagatg
acta 14 <210> SEQ ID NO 180 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 180 gactatgaag atgata 16 <210>
SEQ ID NO 181 <211> LENGTH: 18 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(8) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 181 gatgataagc ctaccaat 18 <210> SEQ ID NO 182
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 182 cgttactctg
aagaag 16 <210> SEQ ID NO 183 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 183 gaagaagaag
agagaccaa 19 <210> SEQ ID NO 184 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 184 gaagaagaga
gaccaaca 18 <210> SEQ ID NO 185 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 185 gaagagagac
caacaaa 17 <210> SEQ ID NO 186 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 186 gaagagaaac
gtcatgtg 18 <210> SEQ ID NO 187 <211> LENGTH: 22
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(12) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 187 gattatagtt
taaaatatgc ca 22 <210> SEQ ID NO 188 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 188 ttaaaatatg ccaca 15 <210>
SEQ ID NO 189 <211> LENGTH: 18 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(8) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 189 gccacagata ttccttca 18 <210> SEQ ID NO 190
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
190 acagatattc cttca 15 <210> SEQ ID NO 191 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: These nucleotides may be
absent <400> SEQUENCE: 191 tcacagaaac agtcat 16 <210>
SEQ ID NO 192 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(5) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 192 aaacagtcat tttca 15 <210> SEQ ID NO 193
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
193 tcaaagagtt catct 15 <210> SEQ ID NO 194 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 194 aaaaccgaac atatg 15 <210>
SEQ ID NO 195 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 195 accgaacata tgtcttc 17 <210> SEQ ID NO 196
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be present <400> SEQUENCE:
196 catatgtctt caagc 15 <210> SEQ ID NO 197 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
present <400> SEQUENCE: 197 ccaagttctg cacaga 16 <210>
SEQ ID NO 198 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be present <400>
SEQUENCE: 198 tgcaaagttt cttcta 16 <210> SEQ ID NO 199
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be present <400> SEQUENCE:
199 atacagactt attgt 15 <210> SEQ ID NO 200 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be present <400> SEQUENCE: 200 cagacttatt
gtgtaga 17 <210> SEQ ID NO 201 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 201 ccaatatgtt tttc 14 <210> SEQ
ID NO 202 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 202 agttcattat
catc 14 <210> SEQ ID NO 203 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 203 caggaagcag attctg 16 <210>
SEQ ID NO 204 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be present <400> SEQUENCE:
204 accctgcaaa tagca 15 <210> SEQ ID NO 205 <211>
LENGTH: 18 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 205 gaaataaaag
aaaagatt 18 <210> SEQ ID NO 206 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 206 ataaaagaaa agat 14 <210> SEQ
ID NO 207 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
207 aaagaaaaga ttggaac 17 <210> SEQ ID NO 208 <211>
LENGTH: 20 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(10) <223> OTHER INFORMATION: These
nucleotides may be present <400> SEQUENCE: 208 aaagaaaaga
ttggaactag 20 <210> SEQ ID NO 209 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 209 gatcctgtga gcgaa 15 <210>
SEQ ID NO 210 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
210 gtgagcgaag ttccag 16 <210> SEQ ID NO 211 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
present <400> SEQUENCE: 211 gttccagcag tgtca 15 <210>
SEQ ID NO 212 <211> LENGTH: 22 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(13) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 212 caccctagaa ccaaatccag ca 22 <210> SEQ ID NO 213
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(6) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
213 agactgcagg gttcta 16 <210> SEQ ID NO 214 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 214 cagggttcta gttt 14 <210> SEQ
ID NO 215 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 215 tctagtttat
cttca 15 <210> SEQ ID NO 216 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 216 ttatcttcag aatca 15 <210>
SEQ ID NO 217 <211> LENGTH: 14 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 4 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
217 gttgaatttt cttc 14 <210> SEQ ID NO 218 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 218 ccctccaaaa gtggt 15 <210>
SEQ ID NO 219 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 219 agtggtgctc agacacc 17 <210> SEQ ID NO 220
<211> LENGTH: 17 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
220 agtccacctg aacacta 17 <210> SEQ ID NO 221 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 221 ccacctgaac actatg 16 <210>
SEQ ID NO 222 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 222 tatgttcagg agaccc 16 <210> SEQ ID NO 223
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(6) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
223 gatagttttg agagtc 16 <210> SEQ ID NO 224 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 224 attgccagct ccgttc 16 <210>
SEQ ID NO 225 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
225 agtggcatta taagcc 16 <210> SEQ ID NO 226 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 226 agccctggac aaacc 15 <210>
SEQ ID NO 227 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
227 cctggacaaa ccatgc 16 <210> SEQ ID NO 228 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 228 atgccaccaa gcaga 15 <210>
SEQ ID NO 229 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
229 aaaaataaag caccta 16 <210> SEQ ID NO 230 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 230 gaaaagagag agag 14 <210> SEQ
ID NO 231 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
231 agagagagtg gacctaa 17 <210> SEQ ID NO 232 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 232 gagagtggac ctaag 15 <210>
SEQ ID NO 233 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 233 gagagtggac ctaagc 16 <210> SEQ ID NO 234
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
234 gagagtggac ctaag 15 <210> SEQ ID NO 235 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 235 gccacggaaa gtac 14 <210> SEQ
ID NO 236 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 236 acggaaagta
ctccag 16 <210> SEQ ID NO 237 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 237 ccagatggat tttc 14 <210> SEQ
ID NO 238 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
238 tcatccagcc tgagtgc 17 <210> SEQ ID NO 239 <211>
LENGTH: 20 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(10) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 239 ttaagaataa
tgcctccagt 20 <210> SEQ ID NO 240 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 240 gaaacagaat
cagagca 17 <210> SEQ ID NO 241 <211> LENGTH: 20
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(10) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 241 tcaaatgaaa
accaagagaa 20 <210> SEQ ID NO 242 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 242 gaaaaccaag agaa 14 <210> SEQ
ID NO 243 <211> LENGTH: 18 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(8) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
243 gagaaagagg cagaaaaa 18 <210> SEQ ID NO 244 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 244 gaatgtatta
tttctg 16 <210> SEQ ID NO 245 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 245 ccagcccaga ctgctt 16 <210>
SEQ ID NO 246 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
246 cagactgctt caaaat 16 <210> SEQ ID NO 247 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 247 ttcaatgata agctc 15 <210>
SEQ ID NO 248 <211> LENGTH: 19 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(9) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 248 aatgattctt tgagttctc 19 <210> SEQ ID NO 249
<211> LENGTH: 19 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(9) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
249 ccagacagag gggcagcaa 19 <210> SEQ ID NO 250 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 250 gaaaatactc cagt 14 <210> SEQ
ID NO 251 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 251 aacaataaag
aaaa 14 <210> SEQ ID NO 252 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 252 gaacctatca aagaga 16 <210>
SEQ ID NO 253 <211> LENGTH: 14 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 4 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
253 cctatcaaag agac 14 <210> SEQ ID NO 254 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 254 gaaccaagta aacct 15 <210>
SEQ ID NO 255 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
255 agctccgcaa tgccaa 16 <210> SEQ ID NO 256 <211>
LENGTH: 23 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(13) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 256 tcatcccttc
ctcgagtaag cac 23 <210> SEQ ID NO 257 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 257 ctaatttatc
aaatggcac 19 <210> SEQ ID NO 258 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: n = C or G <221>
NAME/KEY: misc_feature <222> LOCATION: 7 <223> OTHER
INFORMATION: n = A or n is absent <221> NAME/KEY:
misc_feature <222> LOCATION: 8 <223> OTHER INFORMATION:
n = A or n is absent <400> SEQUENCE: 258 gaagannntt acagcagg
18 <210> SEQ ID NO 259 <211> LENGTH: 18 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: 6
<223> OTHER INFORMATION: n = T or C <221> NAME/KEY:
misc_feature <222> LOCATION: 7 <223> OTHER INFORMATION:
n = T <221> NAME/KEY: misc_feature <222> LOCATION: 8
<223> OTHER INFORMATION: n = A or n is absent <400>
SEQUENCE: 259 cttacnnncc ggagccag 18 <210> SEQ ID NO 260
<211> LENGTH: 25 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
n = C or T <221> NAME/KEY: misc_feature <222> LOCATION:
5 <223> OTHER INFORMATION: n = T <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
n = C or G <221> NAME/KEY: misc_feature <222> LOCATION:
7 <223> OTHER INFORMATION: n = A or C <221> NAME/KEY:
misc_feature <222> LOCATION: 8 <223> OTHER INFORMATION:
n = T or A <221> NAME/KEY: misc_feature <222> LOCATION:
9 <223> OTHER INFORMATION: n = G or n is absent <221>
NAME/KEY: misc_feature <222> LOCATION: (10)...(10)
<223> OTHER INFORMATION: n = C or n is absent <221>
NAME/KEY: misc_feature <222> LOCATION: (11)...(13)
<223> OTHER INFORMATION: n = T or n is absent <221>
NAME/KEY: misc_feature <222> LOCATION: (14)...(15)
<223> OTHER INFORMATION: n = A or n is absent <400>
SEQUENCE: 260 aatnnnnnnn nnnnnggcaa atagg 25 <210> SEQ ID NO
261 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <400> SEQUENCE: 261 ttgcagcttt aa 12
<210> SEQ ID NO 262 <211> LENGTH: 17 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: 5
<223> OTHER INFORMATION: n = G or T <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
n = T or A <221> NAME/KEY: misc_feature <222> LOCATION:
7 <223> OTHER INFORMATION: n = A or n is absent <400>
SEQUENCE: 262 gatgnnntat ggtaaaa 17 <210> SEQ ID NO 263
<211> LENGTH: 21 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 263 tggcgattaa gtcaaattcg
c 21 <210> SEQ ID NO 264 <211> LENGTH: 25 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 264 ccccctagta ccctgacaat gtatt 25 <210> SEQ ID NO
265 <211> LENGTH: 24 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 265 ctgttctgtg
atattatgtg tggt 24 <210> SEQ ID NO 266 <211> LENGTH: 22
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 266 aattgttggc attccagcat tg 22 <210>
SEQ ID NO 267 <211> LENGTH: 38 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 267
gtgcacttac gtgaattcag atgaacgtga tgtagtag 38 <210> SEQ ID NO
268 <211> LENGTH: 28 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 268 tcctcgtact
caacggcttt ctctgaat 28 <210> SEQ ID NO 269 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 269 tccggaacac tagaattctt atttacatac acacttgt
38 <210> SEQ ID NO 270 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 270 cgaataaggt agacggcaac aatgagaa 28 <210> SEQ ID
NO 271 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 271 cggtaaatcg
gagaattcag aggatttaga ggagctaa 38 <210> SEQ ID NO 272
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 272 ctcacgttcg ttacggccat
tgtgatagc 29 <210> SEQ ID NO 273 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 273 ggggaaacag tagaattcca tatggacaga gctgtact
38 <210> SEQ ID NO 274 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 274 tgaagctgtc ggacggcctt tgccctctc 29 <210> SEQ ID
NO 275 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 275 atgggcagtt
atgaattcac tactccctgt agcttgtt 38 <210> SEQ ID NO 276
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 276 tgattggcgc gaacggcact
cagagaaga 29 <210> SEQ ID NO 277 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 277 ctcaagggga ccgaattcgc tggggtcttc tgtgggtc
38 <210> SEQ ID NO 278 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 278 tagggcggcg tgacggccag ccagtggt 28 <210> SEQ ID
NO 279 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 279 gtcttgcatg
tagaattcta gggacgctgc ttttcgtc 38 <210> SEQ ID NO 280
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 280 ctcctagaca tcgggactag
aatgtccac 29 <210> SEQ ID NO 281 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 281 acacaaggca gagaattcca gtcctgaggg tgggggcc
38 <210> SEQ ID NO 282 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 282 ccgtgtttta acgggacaag ctgttcttc 29 <210> SEQ ID
NO 283 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 283 gtagcggagg
ttgaattcta tatgttgtct tggacatt 38 <210> SEQ ID NO 284
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 284 catcagtaga gtgggacgaa
agttctggc 29 <210> SEQ ID NO 285 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 285 atccacgccg cagaattcgt attcatgggc atgtcaaa
38 <210> SEQ ID NO 286 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 286 cttgggacta ttgggaccag tgttcaatc 29 <210> SEQ ID
NO 287 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 287 ccagaaagcc
gtgaattcgt taagccaacc tgactcca 38 <210> SEQ ID NO 288
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 288 tcggggttag tcgggacatc
cagcagccc 29 <210> SEQ ID NO 289 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 289 cgaaggtaat gtgaattcca aaacttagtg ccacaatt
38 <210> SEQ ID NO 290 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 290 ataccgccca acgggacaga tccattgac 29 <210> SEQ ID
NO 291 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 291 agaaacctgt
aagaattcga ttccaaattg ttttttgg 38 <210> SEQ ID NO 292
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 292 cgatcatagg gggggacagg
agagagcac 29 <210> SEQ ID NO 293 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 293 ctactgaggg ctcgtagatc ccaattcctt cccaagct
38 <210> SEQ ID NO 294 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 294 aatcctgctt tagggaccat gctggtgga 29 <210> SEQ ID
NO 295 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 295 caagtgtcct aa
12 <210> SEQ ID NO 296 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 296 cagctgctag aa 12 <210> SEQ ID NO 297
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 297 ggttgagggc aa 12
<210> SEQ ID NO 298 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 298
cacagcgggt aa 12 <210> SEQ ID NO 299 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 299 ttgacttttt aa 12 <210> SEQ ID NO
300 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 300 acagaatggg aa
12 <210> SEQ ID NO 301 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 301 tgcaggtcac aa 12 <210> SEQ ID NO 302
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 302 ttcttcttat aa 12
<210> SEQ ID NO 303 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 303
aggacaacct aa 12 <210> SEQ ID NO 304 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 304 tggtgttcag aa 12 <210> SEQ ID NO
305 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 305 tcagcatatg aa
12 <210> SEQ ID NO 306 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 306 gttgccacac aa 12 <210> SEQ ID NO 307
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 307 cccagctagc aa 12
<210> SEQ ID NO 308 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 308
gggtcactgt aa 12 <210> SEQ ID NO 309 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 309 ttaaataccc aa 12 <210> SEQ ID NO
310 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 310 ttaggaggtt aa
12 <210> SEQ ID NO 311 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 311 acacagaatc aa 12 <210> SEQ ID NO 312
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 312 cgctgaggtc aa 12
<210> SEQ ID NO 313 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 313
aagtagagtc aa 12 <210> SEQ ID NO 314 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 314 cttcccatgg aa 12 <210> SEQ ID NO
315 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 315 ttggttatta aa
12 <210> SEQ ID NO 316 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 316 caacttactg aa 12 <210> SEQ ID NO 317
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 317 cactaagtga aa 12
<210> SEQ ID NO 318 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 318
ctcacctgcc aa 12 <210> SEQ ID NO 319 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 319 atgcatatat aa 12 <210> SEQ ID NO
320 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 320 agagatcagc aa
12 <210> SEQ ID NO 321 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 321 tatatttttc aa 12 <210> SEQ ID NO 322
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 322 cagaaagcag aa 12
<210> SEQ ID NO 323 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 323
gtattgggtt aa 12 <210> SEQ ID NO 324 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 324 ctgacccagg aa 12 <210> SEQ ID NO
325 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 325 cagttttccc aa
12 <210> SEQ ID NO 326 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 326 agggcacagg aa 12 <210> SEQ ID NO 327
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 327 gtatcagagg aa 12
<210> SEQ ID NO 328 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 328
gcatgaaaag aa 12 <210> SEQ ID NO 329 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 329 gatttgacag aa 12 <210> SEQ ID NO
330 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 330 tacagtttac aa
12 <210> SEQ ID NO 331 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 331 tgtgattttt aa 12 <210> SEQ ID NO 332
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 332 ttatgttctc aa 12
<210> SEQ ID NO 333 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 333
caagtacttg aa 12 <210> SEQ ID NO 334 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 334 cttgtgtggc aa 12 <210> SEQ ID NO
335 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 335 agacttctgc aa
12 <210> SEQ ID NO 336 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 336 gttgtctttc aa 12 <210> SEQ ID NO 337
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 337 gggacactcc aa 12
<210> SEQ ID NO 338 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 338
attattattc aa 12 <210> SEQ ID NO 339 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 339 acatgatgac aa 12 <210> SEQ ID NO
340 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 340 tcaattatag aa
12 <210> SEQ ID NO 341 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 341 ctatgggctg aa 12 <210> SEQ ID NO 342
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 342 tgtgtgcctg aa 12
<210> SEQ ID NO 343 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 343
ccatttgttg aa 12 <210> SEQ ID NO 344 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 344 tctccatcaa aa 12 <210> SEQ ID NO
345 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 345 aatgctgaca aa
12 <210> SEQ ID NO 346 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 346 tttcatgtcc aa 12 <210> SEQ ID NO 347
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 347 ggcctcttgg aa 12
<210> SEQ ID NO 348 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 348
tcattttttg aa 12 <210> SEQ ID NO 349 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 349 ggactaccat aa 12 <210> SEQ ID NO
350 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 350 agtcactcag aa
12 <210> SEQ ID NO 351 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 351 ccttggcagg aa 12 <210> SEQ ID NO 352
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 352 tttctggtag aa 12
<210> SEQ ID NO 353 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 353
cccccccccg aa 12 <210> SEQ ID NO 354 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 354 gcccaggcag aa 12 <210> SEQ ID NO
355 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 355 gaatgcgaag aa
12 <210> SEQ ID NO 356 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 356 ttaggtagag aa 12 <210> SEQ ID NO 357
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 357 tgctttggtc aa 12
<210> SEQ ID NO 358 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 358
gcccattaat aa 12 <210> SEQ ID NO 359 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 359 tgagatcttt aa 12 <210> SEQ ID NO
360 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 360 cagtttgttc aa
12 <210> SEQ ID NO 361 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 361 gctgggcaag aa 12 <210> SEQ ID NO 362
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 362 agtcaaagtc aa 12
<210> SEQ ID NO 363 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 363
tctctgcagt aa 12 <210> SEQ ID NO 364 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 364 tgaataactt aa 12 <210> SEQ ID NO
365 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 365 cggttagaaa aa
12 <210> SEQ ID NO 366 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 366 catccctttc aa 12 <210> SEQ ID NO 367
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 367 tctctttctg aa 12
<210> SEQ ID NO 368 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 368
ctcagattgt aa 12 <210> SEQ ID NO 369 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 369 tttgcaccag aa 12 <210> SEQ ID NO
370 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 370 ggttaacatg aa
12 <210> SEQ ID NO 371 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 371 attatcaact aa 12 <210> SEQ ID NO 372
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 372 gccattttgt aa 12
<210> SEQ ID NO 373 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 373
gatctagatg aa 12 <210> SEQ ID NO 374 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 374 ttaatgtatt aa 12 <210> SEQ ID NO
375 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 375 ctagggagac aa
12 <210> SEQ ID NO 376 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 376 tggaggagac aa 12 <210> SEQ ID NO 377
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 377 catcacattt aa 12
<210> SEQ ID NO 378 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 378
ggggtcctgc aa 12 <210> SEQ ID NO 379 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 379 cagttgtgct aa 12 <210> SEQ ID NO
380 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 380 tctgcagcct aa
12 <210> SEQ ID NO 381 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 381 gagtcattta aa 12 <210> SEQ ID NO 382
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 382 tctatggatt aa 12
<210> SEQ ID NO 383 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 383
caaaaagtag aa 12 <210> SEQ ID NO 384 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 384 aatatactcc aa 12 <210> SEQ ID NO
385 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 385 cgtccagcac aa
12 <210> SEQ ID NO 386 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 386 ggatggtgag aa 12 <210> SEQ ID NO 387
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 387 tctcctttgt aa 12
<210> SEQ ID NO 388 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 388
tcgttatttc aa 12 <210> SEQ ID NO 389 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 389 gattttatag aa 12 <210> SEQ ID NO
390 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 390 agacataagc aa
12 <210> SEQ ID NO 391 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 391 ttcacctcac aa 12 <210> SEQ ID NO 392
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 392 ggattgcttg aa 12
<210> SEQ ID NO 393 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 393
actgcatgtg aa 12 <210> SEQ ID NO 394 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 394 tttatcacag aa 12 <210> SEQ ID NO
395 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 395 tcagtaacac aa
12 <210> SEQ ID NO 396 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 396 tacatctttg aa 12 <210> SEQ ID NO 397
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 397 ttgtttcagt aa 12
<210> SEQ ID NO 398 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 398
tatgagcatc aa 12 <210> SEQ ID NO 399 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 399 ctcagcaggc aa 12 <210> SEQ ID NO
400 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 400 acccctgtat aa
12 <210> SEQ ID NO 401 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 401 tctgctcagc aa 12 <210> SEQ ID NO 402
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 402 gttctttttt aa 12
<210> SEQ ID NO 403 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 403
gtgataatcc aa 12 <210> SEQ ID NO 404 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 404 gagccctcag aa 12 <210> SEQ ID NO
405 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 405 tttattggtt aa
12 <210> SEQ ID NO 406 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 406 ggtactgggc aa 12 <210> SEQ ID NO 407
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 407 agtgtttttc aa 12
<210> SEQ ID NO 408 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 408
tgttattggt aa 12 <210> SEQ ID NO 409 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 409 gcgcattcac aa 12 <210> SEQ ID NO
410 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 410 aaacaaaagc aa
12 <210> SEQ ID NO 411 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 411 tatatgatag aa 12 <210> SEQ ID NO 412
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 412 tcccagttcc aa 12
<210> SEQ ID NO 413 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 413
aaagcccata aa 12 <210> SEQ ID NO 414 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 414 tgtcatccac aa 12 <210> SEQ ID NO
415 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 415 ttgtgaatgc aa
12 <210> SEQ ID NO 416 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 416 gtattcatac aa 12 <210> SEQ ID NO 417
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 417 tgacataggg aa 12
<210> SEQ ID NO 418 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 418
agcaaattgc aa 12 <210> SEQ ID NO 419 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 419 agtagatgtt aa 12 <210> SEQ ID NO
420 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 420 aaaagataat aa
12 <210> SEQ ID NO 421 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 421 acctcatggg aa 12 <210> SEQ ID NO 422
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 422 tggtcgacct aa 12
<210> SEQ ID NO 423 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 423
tttgcatggt aa 12 <210> SEQ ID NO 424 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 424 gcggctgccg aa 12 <210> SEQ ID NO
425 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 425 tcaggagtct aa
12 <210> SEQ ID NO 426 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 426 gcctaccagg aa 12 <210> SEQ ID NO 427
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 427 atcttctgtt aa 12
<210> SEQ ID NO 428 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 428
aggtaaggac aa 12 <210> SEQ ID NO 429 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 429 tgctttgagg aa 12 <210> SEQ ID NO
430 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 430 aacagtttta aa
12 <210> SEQ ID NO 431 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 431 ttaaatgttt aa 12 <210> SEQ ID NO 432
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 432 atagaaaatc aa 12
<210> SEQ ID NO 433 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 433
gtgttgtgtt aa 12 <210> SEQ ID NO 434 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 434 gaggacctcg aa 12 <210> SEQ ID NO
435 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 435 agaggctgag aa
12 <210> SEQ ID NO 436 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 436 ggtatttatt aa 12 <210> SEQ ID NO 437
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 437 atttatctgg aa 12
<210> SEQ ID NO 438 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 438
agtgcaaact aa 12 <210> SEQ ID NO 439 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 439 tgaacacctt aa 12 <210> SEQ ID NO
440 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 440 aattttttct aa
12 <210> SEQ ID NO 441 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 441 ttactattat aa 12 <210> SEQ ID NO 442
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 442 tgctatagtg aa 12
<210> SEQ ID NO 443 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 443
tggactatgg aa 12 <210> SEQ ID NO 444 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 444 ctgcagtccg aa 12 <210> SEQ ID NO
445 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 445 gctactgccc aa
12 <210> SEQ ID NO 446 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 446 tcacatggtg aa 12 <210> SEQ ID NO 447
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 447 gtggctctgg aa 12
<210> SEQ ID NO 448 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 448
gaattccatt aa 12 <210> SEQ ID NO 449 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 449 tggggtgtcc aa 12 <210> SEQ ID NO
450 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 450 gcaagctccg aa
12 <210> SEQ ID NO 451 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 451 atgttttttc aa 12 <210> SEQ ID NO 452
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 452 agatctgttg aa 12
<210> SEQ ID NO 453 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 453
aagtgctgtg aa 12 <210> SEQ ID NO 454 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 454 acttttttgg aa 12 <210> SEQ ID NO
455 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 455 aatcggcagg aa
12 <210> SEQ ID NO 456 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 456 ggcatgtcac aa 12 <210> SEQ ID NO 457
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 457 aggaagaaag aa 12
<210> SEQ ID NO 458 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 458
cagtttcacc aa 12 <210> SEQ ID NO 459 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 459 cacagaattt aa 12 <210> SEQ ID NO
460 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 460 aagaataagt aa
12 <210> SEQ ID NO 461 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 461 gggatagtac aa 12 <210> SEQ ID NO 462
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 462 ttcccatgat aa 12
<210> SEQ ID NO 463 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 463
tgattagttg aa 12 <210> SEQ ID NO 464 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 464 gcattcagtg aa 12 <210> SEQ ID NO
465 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 465 agggaatatt aa
12 <210> SEQ ID NO 466 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 466 gaccttaggt aa 12 <210> SEQ ID NO 467
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 467 ttcttttcac aa 12
<210> SEQ ID NO 468 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 468
ccaaactaag aa 12 <210> SEQ ID NO 469 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 469 gtgctcttag aa 12 <210> SEQ ID NO
470 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 470 atgagtttag aa
12 <210> SEQ ID NO 471 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 471 atgagcatag aa 12 <210> SEQ ID NO 472
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 472 gacaaatgag aa 12
<210> SEQ ID NO 473 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 473
aaacccagag aa 12 <210> SEQ ID NO 474 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 474 cctcacacag aa 12 <210> SEQ ID NO
475 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 475 cacactgtgg aa
12 <210> SEQ ID NO 476 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 476 cactgtaccc aa 12 <210> SEQ ID NO 477
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 477 gtagtatttc aa 12
<210> SEQ ID NO 478 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 478
tggatacact aa 12 <210> SEQ ID NO 479 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 479 cccatgattc aa 12 <210> SEQ ID NO
480 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 480 tcataggagg aa
12 <210> SEQ ID NO 481 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 481 aggaaagaga aa 12 <210> SEQ ID NO 482
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 482 atatggtgat aa 12
<210> SEQ ID NO 483 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 483
gatgccatcc aa 12 <210> SEQ ID NO 484 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 484 atactatttc aa 12 <210> SEQ ID NO
485 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 485 gtgtgcatgg aa
12 <210> SEQ ID NO 486 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 486 aggtgttgag aa 12 <210> SEQ ID NO 487
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 487 cagcctgggc aa 12
<210> SEQ ID NO 488 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 488
ggagctctac aa 12 <210> SEQ ID NO 489 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 489 aactaaggtt aa 12 <210> SEQ ID NO
490 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 490 aacttatgtt aa
12 <210> SEQ ID NO 491 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 491 atctcaacag aa 12 <210> SEQ ID NO 492
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 492 taacaatgtg aa 12
<210> SEQ ID NO 493 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 493
aaggatcagg aa 12 <210> SEQ ID NO 494 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 494 ctcaagtctt aa 12 <210> SEQ ID NO
495 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 495 ttagtcatcg
cagaattcta cttctttctg aagtggga 38 <210> SEQ ID NO 496
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 496 ggacagctcg atgggactaa
tgcatactc 29 <210> SEQ ID NO 497 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 497 gtagccactg gtgaattcgt gccatcgcaa aagaataa
38 <210> SEQ ID NO 498 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 498 attagaatga tggggacccc tgtcttccc 29 <210> SEQ ID
NO 499 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 499 acgcatagga
aggaattcat tctgacacgt gtgagata 38 <210> SEQ ID NO 500
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 500 gaaattgacc acgggactgc
acacttttc 29 <210> SEQ ID NO 501 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 501 cggtaaatcg gagaattcaa gttgaggcat gcatccat
38 <210> SEQ ID NO 502 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 502 tcggggctca gcgggaccac agccactcc 29 <210> SEQ ID
NO 503 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 503 tctatgcacc
acgaattcaa tatgtgttca aggacatt 38 <210> SEQ ID NO 504
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 504 tgcttaatcg gtgggacttg
taattgtac 29 <210> SEQ ID NO 505 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 505 cgcgttgtat gcgaattccc tggggtataa agataaga
38 <210> SEQ ID NO 506 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 506 ctcacgggaa ctgggacacc tgaccctgc 29 <210> SEQ ID
NO 507 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 507 gtcttgccgc
ttgaattccc atagaagaat gcgccaaa 38 <210> SEQ ID NO 508
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 508 ttgagtagta cagggacaca
ctaacagac 29 <210> SEQ ID NO 509 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 509 aatactgtag gtgaattctt gcctaagcat tttcccag
38 <210> SEQ ID NO 510 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 510 gtgttgacat tcgggactgt aatcttgac 29 <210> SEQ ID
NO 511 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 511 tctgtagatt
cggaattctt tagagcctgt gcgctgag 38 <210> SEQ ID NO 512
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 512 cgtaccagta cagggacgca
aactgagac 29 <210> SEQ ID NO 513 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 513 gacatgttgg aa 12 <210> SEQ ID NO
514 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 514 acttccagtt aa
12 <210> SEQ ID NO 515 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 515 gtttcctgtt aa 12 <210> SEQ ID NO 516
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 516 cgatgatgac aa 12
<210> SEQ ID NO 517 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 517
gagtagagac aa 12 <210> SEQ ID NO 518 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 518 tcccggatac aa 12 <210> SEQ ID NO
519 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 519 catcctctag aa
12 <210> SEQ ID NO 520 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 520 tattcctgag aa 12 <210> SEQ ID NO 521
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 521 agtttgtttt aa 12
<210> SEQ ID NO 522 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 522
tataaacgat aa 12 <210> SEQ ID NO 523 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 523 tttgaccgat aa 12 <210> SEQ ID NO
524 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 524 tgacaggacc aa
12 <210> SEQ ID NO 525 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 525 ttattcattc aa 12 <210> SEQ ID NO 526
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 526 agtttttcac aa 12
<210> SEQ ID NO 527 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 527
cacctccctg aa 12 <210> SEQ ID NO 528 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 528 ccagattgag aa 12 <210> SEQ ID NO
529 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 529 tgtgtccacc aa
12 <210> SEQ ID NO 530 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 530 cttctattcc aa 12 <210> SEQ ID NO 531
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 531 tcacaatagg aa 12
<210> SEQ ID NO 532 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 532
tacaagtgag aa 12 <210> SEQ ID NO 533 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 533 gagttttcgt aa 12 <210> SEQ ID NO
534 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 534 gtgtgccccc aa
12 <210> SEQ ID NO 535 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 535 gcaccactgc aa 12 <210> SEQ ID NO 536
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 536 gaacacaatg aa 12
<210> SEQ ID NO 537 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 537
tatcctattc aa 12 <210> SEQ ID NO 538 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 538 caaccattat aa 12 <210> SEQ ID NO
539 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 539 tatgctttac aa
12 <210> SEQ ID NO 540 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 540 tttgtttacc aa 12 <210> SEQ ID NO 541
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 541 aggaaattag aa 12
<210> SEQ ID NO 542 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 542
tgttagactt aa 12 <210> SEQ ID NO 543 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 543 tatttggagg aa 12 <210> SEQ ID NO
544 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 544 ggcatttgtc aa
12 <210> SEQ ID NO 545 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 545 atactccagg aa 12 <210> SEQ ID NO 546
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 546 cagcctggac aa 12
<210> SEQ ID NO 547 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 547
ccattgcagt aa 12 <210> SEQ ID NO 548 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 548 aggttctcat aa 12 <210> SEQ ID NO
549 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 549 tgtcatcatt aa
12 <210> SEQ ID NO 550 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 550 tggtatttgc aa 12 <210> SEQ ID NO 551
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 551 tagggtttgt aa 12
<210> SEQ ID NO 552 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 552
ccctaagtag aa 12 <210> SEQ ID NO 553 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 553 gtatttcttt aa 12 <210> SEQ ID NO
554 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 554 gagtcttccc aa
12 <210> SEQ ID NO 555 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 555 caggtagagt aa 12 <210> SEQ ID NO 556
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 556 ataggatgtg aa 12
<210> SEQ ID NO 557 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 557
caatgtgtat aa 12 <210> SEQ ID NO 558 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 558 agagggcatc aa 12 <210> SEQ ID NO
559 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 559 ccagtggtct aa
12 <210> SEQ ID NO 560 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 560 taaacaatag aa 12 <210> SEQ ID NO 561
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 561 gcacactttt aa 12
<210> SEQ ID NO 562 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 562
atggctctgc aa 12 <210> SEQ ID NO 563 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 563 gtcatcttgt aa 12 <210> SEQ ID NO
564 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 564 tgcttcatct aa
12 <210> SEQ ID NO 565 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 565 agaaaggggc aa 12 <210> SEQ ID NO 566
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 566 cttttctttc aa 12
<210> SEQ ID NO 567 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 567
ctactctctc aa 12 <210> SEQ ID NO 568 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 568 acagcattat aa 12 <210> SEQ ID NO
569 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 569 actgctctgg aa
12 <210> SEQ ID NO 570 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 570 gcagaggcac aa 12 <210> SEQ ID NO 571
<211> LENGTH: 38 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 571 ctccgtggta tggaattcca
ctcaaatctt cattcaga 38 <210> SEQ ID NO 572 <211>
LENGTH: 29 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 572 acgtcgggtt acgggacacc tgattcctc 29
<210> SEQ ID NO 573 <211> LENGTH: 38 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 573
taccattggt ttgaattctt gtttcctgtt aaccatgc 38 <210> SEQ ID NO
574 <211> LENGTH: 29 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 574 gccgagttct
acgggacaga aaagggagc 29 <210> SEQ ID NO 575 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 575 tgcagtgatt tcgaattcga gacaatgctg cccagtca
38 <210> SEQ ID NO 576 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 576 tctaaattct ctgggaccat tccttcaac 29 <210> SEQ ID
NO 577 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 577 actaacagca
ctgaattcca tgctcttgga ctttccat 38 <210> SEQ ID NO 578
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 578 tcccctaacg ttgggacaca
gaatactac 29 <210> SEQ ID NO 579 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 579 gtcgacgatg gcgaattcct gccactcatt cagttagc
38 <210> SEQ ID NO 580 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 580 gaacggccca cagggacctg gcataactc 29 <210> SEQ ID
NO 581 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 581 tcatggtagc
aggaattctg ctttgaccga taaggaga 38 <210> SEQ ID NO 582
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 582 actgtgggat tcgggactgt
ctactaccc 29 <210> SEQ ID NO 583 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 583 acctctcggc cggaattcgg aaaagtgtac agatcatt
38 <210> SEQ ID NO 584 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 584 gccggatacg aagggacggc tcgtgactc 29 <210> SEQ ID
NO 585 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 585 ccgtagacta
aagaattccc tgatgtcagg ctgtcacc 38 <210> SEQ ID NO 586
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 586 atcggatcag tcgggacggt
gtctttgcc 29 <210> SEQ ID NO 587 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 587 gcataggcgg gagaattccc tgtgtccacc aaagtcgg
38 <210> SEQ ID NO 588 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 588 cccacatagg gcgggacaaa gagctgaac 29 <210> SEQ ID
NO 589 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 589 ggcttgccga
gcgaattcta ggaaagatac ggaatcaa 38 <210> SEQ ID NO 590
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 590 taaccctcat acgggacttt
catggaagc 29 <210> SEQ ID NO 591 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 591 atgagcaccc gggaattctg attggagtct aggccaaa
38 <210> SEQ ID NO 592 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 592 tgctcacctt ctgggacgtg gctggtctc 29 <210> SEQ ID
NO 593 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 593 accgtctgcc
acgaattctg gaaaacatgc agtctggt 38 <210> SEQ ID NO 594
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 594 tacacgggag gcgggacagg
gtgattaac 29 <210> SEQ ID NO 595 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 595 cttaaagcta acgaattcag agctgtatga agatgctt
38 <210> SEQ ID NO 596 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 596 aacgctaaag gggggacaac ataattggc 29 <210> SEQ ID
NO 597 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 597 ttgtaagaac
gagaattctg caacctgtct ttattgaa 38 <210> SEQ ID NO 598
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 598 cttcaccact ttgggacact
gaagccaac 29 <210> SEQ ID NO 599 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 599 aaccattgat ttgaattcga aatgtccacc aaagttca
38 <210> SEQ ID NO 600 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 600 tgtctagttc cagggacgct gttacttac 29 <210> SEQ ID
NO 601 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 601 cgaaggtaat
gtgaattctg ccacaattaa gacttgga 38 <210> SEQ ID NO 602
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 602 ataccggttt tcgggacaga
tccattgac 29 <210> SEQ ID NO 603 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 603 cctgaaatcc acgaattcca ccctggcctc ccagtgca
38 <210> SEQ ID NO 604 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 604 tagatggtag gtgggacagg actggcttc 29 <210> SEQ ID
NO 605 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 605 gcatatctta
gcgaattcct gtgactaata cagagtgc 38 <210> SEQ ID NO 606
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 606 ccaaatatgg tagggacgtg
tgaacactc 29 <210> SEQ ID NO 607 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 607 tgccgctaca gggaattcat atggcagata ttcctgaa
38 <210> SEQ ID NO 608 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 608 acgttgcgga ccgggacttc cacagagcc 29 <210> SEQ ID
NO 609 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 609 cttcgcccaa
tggaattcgg tacaggggta tgccttat 38 <210> SEQ ID NO 610
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 610 tgcacttctg ccgggaccag
aggagaaac 29 <210> SEQ ID NO 611 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 611 tgtgggtatt ctgaattcca caaaatggac taacacgc
38 <210> SEQ ID NO 612 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 612 acgtcgttca gtgggacatt aaaaggctc 29 <210> SEQ ID
NO 613 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 613 ggttatgtgt
cagaattctg aaactagttt ggaagtac 38 <210> SEQ ID NO 614
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 614 gcctcagttt cggggacagt
tctgaggac 29 <210> SEQ ID NO 615 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 615 tgtaacacgg ccgaattcct catttgtatg aaataggt
38 <210> SEQ ID NO 616 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 616 aatctaactt gagggaccgg cacacacac 29 <210> SEQ ID
NO 617 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 617 agtgtcccct
tagaattcgc agagacacca cagtgtgc 38 <210> SEQ ID NO 618
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 618 tttgctacag tcgggaccct
tgtgtgctc 29 <210> SEQ ID NO 619 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 619 agcacatcac tagaattcaa taccatgtgt gagctcaa
38 <210> SEQ ID NO 620 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 620 aatcctgctt ccgggaccta actttgaac 29 <210> SEQ ID
NO 621 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 621 tttcattttc
tggaattcct ctaatgattt tctggagc 38 <210> SEQ ID NO 622
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 622 cgtcgccgcg tagggacttt
ttcttccac 29 <210> SEQ ID NO 623 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 623 ttacttaatc ctgaattcga gaaaagccat gttgataa
38 <210> SEQ ID NO 624 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 624 tcatgggtcg ctgggacttt gccctctgc 29 <210> SEQ ID
NO 625 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 625 actaacagca
ctgaattcat tttactataa tctgctac 38 <210> SEQ ID NO 626
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 626 gttagccgag aagggactgt
ctgtgaagc 29 <210> SEQ ID NO 627 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 627 aaatatgcag cggaattcgt aagtgaccta ttaataac
38 <210> SEQ ID NO 628 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 628 gcgatggtta cggggacagc caggcaacc 29 <210> SEQ ID
NO 629 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 629 ccagctggta
gaactt 16 <210> SEQ ID NO 630 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 630 cccaatagac ctatag 16 <210> SEQ ID
NO 631 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 631 tagcagaatc
tctcat 16 <210> SEQ ID NO 632 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 632 agagtatctc atttgtt 17 <210> SEQ ID
NO 633 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 633 aggaaattgt
gaagta 16 <210> SEQ ID NO 634 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 634 taactcactc actatc 16 <210> SEQ ID
NO 635 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 635 ctgctgagtc
atagtc 16 <210> SEQ ID NO 636 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 636 tgttctttga atcaac 16 <210> SEQ ID
NO 637 <211> LENGTH: 47 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 637 cagtaatacg
actcactata ggggtcagga ttagcagaat ctctcat 47 <210> SEQ ID NO
638 <211> LENGTH: 27 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 638 gcattctatg
agagtatctc atttgtt 27 <210> SEQ ID NO 639 <211> LENGTH:
47 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 639 cagtaatacg actcactata ggggtcagga
agagtatctc atttgtt 47 <210> SEQ ID NO 640 <211> LENGTH:
26 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 640 gcattctatg tagcagaatc tctcat 26
<210> SEQ ID NO 641 <211> LENGTH: 17 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(17) <223> OTHER
INFORMATION: n = T or A <400> SEQUENCE: 641 ccgcatanct
cagcaca 17 <210> SEQ ID NO 642 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<221> NAME/KEY: misc_feature <222> LOCATION: (1)...(17)
<223> OTHER INFORMATION: n = A or T <400> SEQUENCE: 642
tgtgctgagn tatgcgg 17 <210> SEQ ID NO 643 <211> LENGTH:
16 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 643 aggaaattgt gaagta 16 <210> SEQ ID
NO 644 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 644 taactcactc
actatc 16 <210> SEQ ID NO 645 <211> LENGTH: 25
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 645 ctttttatgc ctttccactc atcca 25
<210> SEQ ID NO 646 <211> LENGTH: 25 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 646
ctttttatgc ctttccactc atcca 25 <210> SEQ ID NO 647
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <400> SEQUENCE: 647 gaaaaatacg gaaaggtgag
taggtttcc 29 <210> SEQ ID NO 648 <211> LENGTH: 26
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 648 ctttttatgc ctttccactc atccaa 26
<210> SEQ ID NO 649 <211> LENGTH: 29 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <400> SEQUENCE: 649
gaaaaatacg gaaaggtgag taggtttcc 29 <210> SEQ ID NO 650
<211> LENGTH: 25 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 650 ctttttatgc ctttccactc
atcca 25 <210> SEQ ID NO 651 <211> LENGTH: 29
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<400> SEQUENCE: 651 gaaaaatacg gaaaggtgag taggtctcc 29
<210> SEQ ID NO 652 <211> LENGTH: 27 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 652
ctttttatgc ctttccactc atccaga 27 <210> SEQ ID NO 653
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <400> SEQUENCE: 653 gaaaaatacg gaaaggtgag
taggtctcc 29 <210> SEQ ID NO 654 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 654 atcacactgg ggatc 15 <210> SEQ ID NO
655 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 655 ctaaacctat
gactc 15 <210> SEQ ID NO 656 <211> LENGTH: 21
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 656 ttcacagagg atatcttaat a 21 <210>
SEQ ID NO 657 <211> LENGTH: 22 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 657
agtcttgtaa tacgacagtc tt 22 <210> SEQ ID NO 658 <211>
LENGTH: 21 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 658 ccatatcaat cagtactctt g 21 <210>
SEQ ID NO 659 <211> LENGTH: 23 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Oligonucleotide probe <400>
SEQUENCE: 659 cctttccact catccaaagg ttg 23 <210> SEQ ID NO
660 <211> LENGTH: 23 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Oligonucleotide probe <400> SEQUENCE: 660
cctttccact catccagagg ttg 23 <210> SEQ ID NO 661 <211>
LENGTH: 20 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION:
Oligonucleotide probe <400> SEQUENCE: 661 catgcagata
taccgcatat 20 <210> SEQ ID NO 662 <211> LENGTH: 20
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Oligonucleotide
probe <400> SEQUENCE: 662 catgcagata taccgcataa 20
<210> SEQ ID NO 663 <211> LENGTH: 23 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Oligonucleotide probe <400>
SEQUENCE: 663 ggtagcatct ctcagcacaa gag 23 <210> SEQ ID NO
664 <211> LENGTH: 23 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Oligonucleotide probe <400> SEQUENCE: 664
ggtagcatca ctcagcacaa gag 23 <210> SEQ ID NO 665 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 665 tagcagaatc tctcat 16 <210> SEQ ID
NO 666 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 666 agagtatctc
atttgtt 17 <210> SEQ ID NO 667 <211> LENGTH: 125
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(125) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 667 gccaagtata attttcctga taaaggctgg
gctgcnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnncaattt gcactartga agtttaccta 120 acaat 125
<210> SEQ ID NO 668 <211> LENGTH: 125 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(125) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 668 attgttaggt
aaacttcayt agtgcaaatt gnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn gcagcccagc ctttatcagg aaaattatac
120 ttggc 125 <210> SEQ ID NO 669 <211> LENGTH: 140
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(140) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 669 ctatatgtga aagcagaagt tgtctgataa
tcnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnccaagg 120 sctttactcg
atgatagctg 140 <210> SEQ ID NO 670 <211> LENGTH: 140
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(140) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 670 cagctatcat cgagtaaags ccttggnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnga ttatcagaca 120 acttctgctt
tcacatatag 140 <210> SEQ ID NO 671 <211> LENGTH: 141
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(141) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 671 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tartgaagtt 120 tacctagtcc
ccagatttaa g 141 <210> SEQ ID NO 672 <211> LENGTH: 141
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(141) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 672 cttaaatcag gggactaggt aaacttcayt
agtgcaaatt gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn cagcctttat caggaaaatt 120 atagaattcg
catacggtta t 141 <210> SEQ ID NO 673 <211> LENGTH: 141
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(141) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 673 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tartgaagtt 120 tacctagccg
tcagatttaa g 141 <210> SEQ ID NO 674 <211> LENGTH: 141
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(141) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 674 cttaaatcag acggctaggt aaacttcayt
agtgcaaatt gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn cagcctttat caggaaaatt 120 atagaattcg
catacggtta t 141 <210> SEQ ID NO 675 <211> LENGTH: 161
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(161) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 675 aagtttagat cagaattcgt gaaagcagaa
gttgtctgat aatcnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 120 nnnnnnccaa
ggsctttact cgatgagtcc cttatcgtga t 161 <210> SEQ ID NO 676
<211> LENGTH: 161 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR Product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(161) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 676 atcacgataa gggactcatc
gagtaaagsc cttggnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnngat 120
tatcagacaa cttctgcttt cacgaattct gatctaaact t 161 <210> SEQ
ID NO 677 <211> LENGTH: 161 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR Product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(161) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 677 aagtttagat cagaattcgt
gaaagcagaa gttgtctgat aatcnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 120
nnnnnnccaa ggsctttact cgatgagccg tttatcgtga t 161 <210> SEQ
ID NO 678 <211> LENGTH: 161 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR Product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(161) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 678 atcacgataa acggctcatc
gagtaaagsc cttggnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnngat 120
tatcagacaa cttctgcttt cacgaattct gatctaaact t 161 <210> SEQ
ID NO 679 <211> LENGTH: 112 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(112) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 679 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac ta 112 <210> SEQ ID NO 680
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 680 rtgaagttta cctagtcccc agatttaag 29
<210> SEQ ID NO 681 <211> LENGTH: 116 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(116) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 681 tcaytagtgc aaattgnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnncagcc
tttatcagga aaattataga attcgcatac ggttat 116 <210> SEQ ID NO
682 <211> LENGTH: 25 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 682 cttaaatcag gggactaggt aaact 25
<210> SEQ ID NO 683 <211> LENGTH: 29 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 683 rtgaagttta cctagccgtc agatttaag
29 <210> SEQ ID NO 684 <211> LENGTH: 114 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: PCR product after
restriction enzyme digestion <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(114) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 684 aytagtgcaa attgnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnncagcctt tatcaggaaa attatagaat tcgcatacgg ttat 114 <210>
SEQ ID NO 685 <211> LENGTH: 27 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 685 cttaaatcag acggctaggt aaacttc
27 <210> SEQ ID NO 686 <211> LENGTH: 132 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: PCR product after
restriction enzyme digestion <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(132) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 686 aagtttagat cagaattcgt
gaaagcagaa gttgtctgat aatcnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 120
nnnnnnccaa gg 132 <210> SEQ ID NO 687 <211> LENGTH: 29
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR product
after restriction enzyme digestion <400> SEQUENCE: 687
sctttactcg atgagtccct tatcgtgat 29 <210> SEQ ID NO 688
<211> LENGTH: 136 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(136) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 688 aagsccttgg nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nngattatca gacaacttct gctttcacga 120 attctgatct aaactt
136 <210> SEQ ID NO 689 <211> LENGTH: 25 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: PCR product after
restriction enzyme digestion <400> SEQUENCE: 689 atcacgataa
gggactcatc gagta 25 <210> SEQ ID NO 690 <211> LENGTH:
29 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR product
after restriction enzyme digestion <400> SEQUENCE: 690
sctttactcg atgagccgtt tatcgtgat 29 <210> SEQ ID NO 691
<211> LENGTH: 134 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(134) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 691 gsccttggnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn gattatcaga caacttctgc tttcacgaat 120 tctgatctaa actt 134
<210> SEQ ID NO 692 <211> LENGTH: 27 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 692 atcacgataa acggctcatc gagtaaa
27 <210> SEQ ID NO 693 <211> LENGTH: 113 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Polymerase extension
product <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(113) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 693 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tar 113 <210> SEQ ID NO 694
<211> LENGTH: 116 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Polymerase extension product <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(116) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 694 tcaytagtgc
aaattgnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnncagcc tttatcagga aaattataga attcgcatac ggttat 116
<210> SEQ ID NO 695 <211> LENGTH: 114 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Polymerase extension product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(114) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 695 aytagtgcaa attgnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnncagcctt
tatcaggaaa attatagaat tcgcatacgg ttat 114 <210> SEQ ID NO 696
<211> LENGTH: 133 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Polymerase extension product <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(133) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 696 aagtttagat
cagaattcgt gaaagcagaa gttgtctgat aatcnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn
120 nnnnnnccaa ggs 133 <210> SEQ ID NO 697 <211>
LENGTH: 136 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION:
Polymerase extension product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(136) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 697 aagsccttgg nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nngattatca gacaacttct gctttcacga 120
attctgatct aaactt 136 <210> SEQ ID NO 698 <211> LENGTH:
134 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Polymerase
extension product <221> NAME/KEY: misc_feature <222>
LOCATION: (1)...(134) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 698 gsccttggnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn gattatcaga caacttctgc tttcacgaat 120 tctgatctaa actt 134
<210> SEQ ID NO 699 <211> LENGTH: 114 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Polymerase extension product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(114) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 699 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tart 114 <210> SEQ ID NO 700
<211> LENGTH: 115 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Polymerase extension product <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(115) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 700 ataaccgtat
gcgaattcta taattttcct gataaaggct gnnnnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn caatttgcac tartg 115
<210> SEQ ID NO 701 <211> LENGTH: 116 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Polymerase extension product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(116) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 701 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tartga 116 <210> SEQ ID NO
702 <211> LENGTH: 13 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 702 ataaccgtat gcg 13 <210> SEQ ID NO
703 <211> LENGTH: 100 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(100) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 703 aattctataa ttttcctgat aaaggctgnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnncaa tttgcactar 100 <210> SEQ ID NO 704 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: PCR
product after restriction enzyme digestion <400> SEQUENCE:
704 aattcgcata cggttat 17 <210> SEQ ID NO 705 <211>
LENGTH: 99 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: PCR
product after restriction enzyme digestion <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(99) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 705 tcaytagtgc
aaattgnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnncagcc tttatcagga aaattatag 99 <210> SEQ ID NO
706 <211> LENGTH: 97 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION: (1)...(97)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
706 aytagtgcaa attgnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn 60 nnnnnnnnnn nnncagcctt tatcaggaaa attatag 97
<210> SEQ ID NO 707 <211> LENGTH: 13 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 707 aagtttagat cag 13 <210>
SEQ ID NO 708 <211> LENGTH: 120 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(120) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 708 aattcgtgaa agcagaagtt gtctgataat
cnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnccaaggs 120 <210> SEQ ID
NO 709 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 709 aattctgatc taaactt 17 <210> SEQ ID
NO 710 <211> LENGTH: 119 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(119) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 710 aagsccttgg nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nngattatca gacaacttct gctttcacg 119 <210> SEQ ID
NO 711 <211> LENGTH: 117 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION: (9)...(90)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
711 gsccttggnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn gattatcaga
caacttctgc tttcacg 117
1 SEQUENCE LISTING <160> NUMBER OF SEQ ID NOS: 711
<210> SEQ ID NO 1 <211> LENGTH: 15 <212> TYPE:
DNA <213> ORGANISM: Unknown <220> FEATURE: <223>
OTHER INFORMATION: Restriction site <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(15) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 1 gggacnnnnn
nnnnn 15 <210> SEQ ID NO 2 <211> LENGTH: 19 <212>
TYPE: DNA <213> ORGANISM: Unknown <220> FEATURE:
<223> OTHER INFORMATION: Restriction site <221>
NAME/KEY: misc_feature <222> LOCATION: (1)...(14) <223>
OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE: 2
nnnnnnnnnn nnnngtccc 19 <210> SEQ ID NO 3 <211> LENGTH:
21 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 3 ggaaattcca tgatgcgtgg g 21 <210> SEQ
ID NO 4 <211> LENGTH: 23 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (19)...(21) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 4 ggaaattcca
tgatgcgtnn nac 23 <210> SEQ ID NO 5 <211> LENGTH: 21
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 5 ggaaattcca tgatgcgtac c 21 <210> SEQ
ID NO 6 <211> LENGTH: 25 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (22)...(23) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 6 ggaaattcca
tgatgcgtac cnngg 25 <210> SEQ ID NO 7 <211> LENGTH: 11
<212> TYPE: DNA <213> ORGANISM: Unknown <220>
FEATURE: <223> OTHER INFORMATION: Restriction site
<221> NAME/KEY: misc_feature <222> LOCATION: (4)...(8)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
7 cctnnnnnag g 11 <210> SEQ ID NO 8 <211> LENGTH: 25
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (20)...(23) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 8 ggaaattcca tgatgcgtan nnngg 25 <210>
SEQ ID NO 9 <211> LENGTH: 38 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 9
tagaatagca ctgaattcag gaatacaatc attgtcac 38 <210> SEQ ID NO
10 <211> LENGTH: 28 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 10 atcacgataa
acggccaaac tcaggtta 28 <210> SEQ ID NO 11 <211> LENGTH:
38 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 11 aagtttagat cagaattcgt gaaagcagaa gttgtctg
38 <210> SEQ ID NO 12 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 12 tctccaacta acggctcatc gagtaaag 28 <210> SEQ ID
NO 13 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 13 atgactagct
atgaattcgt tcaaggtaga aaatggaa 38 <210> SEQ ID NO 14
<211> LENGTH: 28 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 14 gagaattaga acggcccaaa
tcccactc 28 <210> SEQ ID NO 15 <211> LENGTH: 37
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 15 ttacaatgca tgaattcatc ttggtctctc aaagtgc
37 <210> SEQ ID NO 16 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 16 tggaccataa acggccaaaa actgtaag 28 <210> SEQ ID
NO 17 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 17 ataaccgtat
gcgaattcta taattttcct gataaagg 38 <210> SEQ ID NO 18
<211> LENGTH: 28 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 18 cttaaatcag gggactaggt
aaacttca 28 <210> SEQ ID NO 19 <211> LENGTH: 28
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 19 cttaaatcag acggctaggt aaacttca 28
<210> SEQ ID NO 20
<211> LENGTH: 28 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 20 tctccaacta gggactcatc
gagtaaag 28 <210> SEQ ID NO 21 <211> LENGTH: 37
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 21 aacgccgggc gagaattcag tttttcaact tgcaagg
37 <210> SEQ ID NO 22 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 22 ctacacatat ctgggacgtt ggccatcc 28 <210> SEQ ID
NO 23 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 23 taccttttga
tcgaattcaa ggccaaaaat attaagtt 38 <210> SEQ ID NO 24
<211> LENGTH: 28 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 24 tcgaacttta acggccttag
agtagaga 28 <210> SEQ ID NO 25 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 25 cgatttcgat aagaattcaa aagcagttct tagttcag
38 <210> SEQ ID NO 26 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 26 tgcgaatctt acggctgcat cacattca 28 <210> SEQ ID
NO 27 <211> LENGTH: 23 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (3)...(5) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 27 gtnnnacgca
tcatggaatt tcc 23 <210> SEQ ID NO 28 <211> LENGTH: 25
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (3)...(4) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 28 ccnnggtacg catcatggaa tttcc 25 <210>
SEQ ID NO 29 <211> LENGTH: 25 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (3)...(6) <223>
OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE: 29
ccnnnntacg catcatggaa tttcc 25 <210> SEQ ID NO 30 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 30 gggctagtct ccgaattcca cctatcctac caaatgtc
38 <210> SEQ ID NO 31 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 31 tagctgtagt tagggactgt tctgagcac 29 <210> SEQ ID
NO 32 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 32 cgaatgcaag
gcgaattcgt tagtaataac acagtgca 38 <210> SEQ ID NO 33
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 33 aagactggat ccgggaccat
gtagaatac 29 <210> SEQ ID NO 34 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 34 tctaaccatt gcgaattcag ggcaaggggg gtgagatc
38 <210> SEQ ID NO 35 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 35 tgacttggat ccgggacaac gactcatcc 29 <210> SEQ ID
NO 36 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 36 acccaggcgc
cagaattctt tagataaagc tgaaggga 38 <210> SEQ ID NO 37
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 37 gttacgggat ccgggactcc
atattgatc 29 <210> SEQ ID NO 38 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 38 cgttggcttg aggaattcga ccaaaagagc caagagaa
38 <210> SEQ ID NO 39 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 39 aaaaagggat ccgggacctt gactaggac 29 <210> SEQ ID
NO 40 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer
<400> SEQUENCE: 40 acttgattcc gtgaattcgt tatcaataaa tcttacat
38 <210> SEQ ID NO 41 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 41 caagttggat ccgggaccca gggctaacc 29 <210> SEQ ID
NO 42 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 42 gtgcaaaggc
ctgaattccc aggcacaaag ctgttgaa 38 <210> SEQ ID NO 43
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 43 tgaagcgaac tagggactca
ggtggactt 29 <210> SEQ ID NO 44 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 44 gattccgtaa acgaattcag ttcattatca tctttgtc
38 <210> SEQ ID NO 45 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 45 ccattgttaa gcgggacttc tgctatttg 29 <210> SEQ ID
NO 46 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <400> SEQUENCE: 46 cccaaaagtc cacctga
17 <210> SEQ ID NO 47 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <400> SEQUENCE:
47 tcaggtggac ttttggg 17 <210> SEQ ID NO 48 <211>
LENGTH: 18 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<400> SEQUENCE: 48 accctgcaaa tagcagaa 18 <210> SEQ ID
NO 49 <211> LENGTH: 18 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <400> SEQUENCE: 49 ttctgctatt tgcagggt
18 <210> SEQ ID NO 50 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <400> SEQUENCE:
50 acccgcaaat agcagaa 17 <210> SEQ ID NO 51 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<400> SEQUENCE: 51 ttctgctatt tgcgggt 17 <210> SEQ ID
NO 52 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
52 ttagatagca gtaattt 17 <210> SEQ ID NO 53 <211>
LENGTH: 23 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(13) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 53 ggaagccggg
aaggatctgt atc 23 <210> SEQ ID NO 54 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 54 gagaaagaga ggtaa 15 <210> SEQ
ID NO 55 <211> LENGTH: 19 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(8) <223> OTHER
INFORMATION: These nucleotides may be absent <221> NAME/KEY:
misc_feature <222> LOCATION: (10)...(19) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
55 aaagagaggt aacttttct 19 <210> SEQ ID NO 56 <211>
LENGTH: 18 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 56 aaagagaggt
aacttttc 18 <210> SEQ ID NO 57 <211> LENGTH: 23
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 11 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 57 ttttaaaaaa aaaaaatagg tca 23
<210> SEQ ID NO 58 <211> LENGTH: 22 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION:
(11)...(12) <223> OTHER INFORMATION: These nucleotides may be
absent <400> SEQUENCE: 58 aaaataggtc attgcttctt gc 22
<210> SEQ ID NO 59 <211> LENGTH: 16 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (5)...(6)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 59 gacaaagaag aaaagg 16 <210> SEQ ID NO
60 <211> LENGTH: 19 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(9) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
60
gacaaagaag aaaaggaaa 19 <210> SEQ ID NO 61 <211>
LENGTH: 25 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (11)...(15) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 61 aggaaaaaga
ctggtattac gctca 25 <210> SEQ ID NO 62 <211> LENGTH: 24
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (11)...(14) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 62 aaaagaatag
atagtcttcc ttta 24 <210> SEQ ID NO 63 <211> LENGTH: 21
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 11 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 63 agatagtctt cctttaactg a 21
<210> SEQ ID NO 64 <211> LENGTH: 19 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(9)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 64 tccttacaaa cagatatga 19 <210> SEQ ID
NO 65 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 65 accagaaggc
aatt 14 <210> SEQ ID NO 66 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(7)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 66 atcagagttg cgatgga 17 <210> SEQ ID
NO 67 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 67 cgagcacagg
taagtt 16 <210> SEQ ID NO 68 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 68 cactctgcac ctcga
15 <210> SEQ ID NO 69 <211> LENGTH: 14 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: 4
<223> OTHER INFORMATION: This nucleotide may be absent
<400> SEQUENCE: 69 gatatgtcgc gaac 14 <210> SEQ ID NO
70 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
70 aaagactctg tattgtt 17 <210> SEQ ID NO 71 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 71 gacaagagag gcagg
15 <210> SEQ ID NO 72 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(7)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 72 catgaaccag gcatgga 17 <210> SEQ ID
NO 73 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 73 gaaccaggca
tggacc 16 <210> SEQ ID NO 74 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 74 aatccaagta
tgttctct 18 <210> SEQ ID NO 75 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 75 gctcctgttg aacatc 16 <210>
SEQ ID NO 76 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE: 76
aaactttcat ttgatg 16 <210> SEQ ID NO 77 <211> LENGTH:
19 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 77 aaactttcat
ttgatgaag 19 <210> SEQ ID NO 78 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent
<400> SEQUENCE: 78 ctacaggcca ttgc 14 <210> SEQ ID NO
79 <211> LENGTH: 21 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 11 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE: 79
taaattaggg ggactacagg c 21 <210> SEQ ID NO 80 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 80 ttattgcaag tggac 15 <210> SEQ
ID NO 81 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 81 tacgggctta
ctaat 15 <210> SEQ ID NO 82 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 82 agtattacac taagac 16 <210>
SEQ ID NO 83 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 83 attacactaa gacgata 17 <210> SEQ ID NO 84
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 84 ctaagacgat
atgc 14 <210> SEQ ID NO 85 <211> LENGTH: 17 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(7)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 85 tgctctatga aaggctg 17 <210> SEQ ID
NO 86 <211> LENGTH: 25 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(15) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
86 atgagagcac ttgtggccca actaa 25 <210> SEQ ID NO 87
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 87 gacttacagc
aggtac 16 <210> SEQ ID NO 88 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 88 aaaaagacgt
tgcgaga 17 <210> SEQ ID NO 89 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 89 gttggaagtg
tgaaagcat 19 <210> SEQ ID NO 90 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 90 aaagcattga tggaat 16 <210>
SEQ ID NO 91 <211> LENGTH: 18 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(8) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 91 ttagaagtta aaaaggta 18 <210> SEQ ID NO 92
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 92 accctcaaaa
gcgtat 16 <210> SEQ ID NO 93 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 93 gccttatgga atttg 15 <210> SEQ
ID NO 94 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 94 gctgtagatg
gtgc 14 <210> SEQ ID NO 95 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (4)...(19)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 95 gttggcactc ttacttaccg gagccagac 29
<210> SEQ ID NO 96 <211> LENGTH: 19 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(9)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 96 cttacttacc ggagccaga 19 <210> SEQ ID
NO 97 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 97 acttaccgga
gccag 15 <210> SEQ ID NO 98 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 98 agccagacaa acact
15 <210> SEQ ID NO 99 <211> LENGTH: 18 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (5)...(8)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 99 agccagacaa acacttta 18 <210> SEQ ID
NO 100 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
100 acaaacactt tagccat 17 <210> SEQ ID NO 101 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 101 ttagccatta ttgaaa 16 <210>
SEQ ID NO 102 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
102 ggaggtggga tatta 15 <210> SEQ ID NO 103 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 103 atattacgga atgtg 15 <210>
SEQ ID NO 104 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
104 ttacggaatg tgtcca 16 <210> SEQ ID NO 105 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 105 agagagaaca
actgt 15 <210> SEQ ID NO 106 <211> LENGTH: 34
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (11)...(24) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 106 tatttcaggc
aaatcctaag agagaacaac tgtc 34 <210> SEQ ID NO 107 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 107 aactgtctac
aaactt 16 <210> SEQ ID NO 108 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 108 ttattacaac
actta 15 <210> SEQ ID NO 109 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 109 cacttaaaat
ctcat 15 <210> SEQ ID NO 110 <211> LENGTH: 24
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(14) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 110 agtttgacaa
tagtcagtaa tgca 24 <210> SEQ ID NO 111 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 111 cacttatcag aaactt 16 <210>
SEQ ID NO 112 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
112 ttatcagaaa ctttt 15 <210> SEQ ID NO 113 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 113 tcagaaactt ttgaca 16 <210>
SEQ ID NO 114 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5
<223> OTHER INFORMATION: This nucleotide may be absent
<400> SEQUENCE: 114 agtcccaagg catct 15 <210> SEQ ID NO
115 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 115 aagcaaagtc
tctat 15 <210> SEQ ID NO 116 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 116 aagcaaagtc
tctatgg 17 <210> SEQ ID NO 117 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: These nucleotides may be
absent <400> SEQUENCE: 117 caaagtctct atggt 15 <210>
SEQ ID NO 118 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 118 gattatgttt ttgaca 16 <210> SEQ ID NO 119
<211> LENGTH: 25 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(15) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
119 gacaccaatc gacatgatga taata 25 <210> SEQ ID NO 120
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(6) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
120 cgacatgatg ataata 16 <210> SEQ ID NO 121 <211>
LENGTH: 18 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 121 tcagacaatt
ttaatact 18 <210> SEQ ID NO 122 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 122 tatttgaata ctac 14 <210> SEQ
ID NO 123 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 123 aatactacag
tgtta 15 <210> SEQ ID NO 124 <211> LENGTH: 28
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(18) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 124 gtgttaccca
gctcctcttc atcaagag 28 <210> SEQ ID NO 125 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 125 agctcctctt catcaa 16 <210>
SEQ ID NO 126 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
126 tcatcaagag gaagc 15 <210> SEQ ID NO 127 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 127 aaagatagaa
gtttgga 17 <210> SEQ ID NO 128 <211> LENGTH: 21
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(11) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 128 aaagatagaa
gtttggagag a 21 <210> SEQ ID NO 129 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 129 gaacgcggaa ttggt 15 <210>
SEQ ID NO 130 <211> LENGTH: 19 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(9) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 130 cgcggaattg gtctaggca 19 <210> SEQ ID NO 131
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 131 cgcggaattg
gtcta 15 <210> SEQ ID NO 132 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature
<222> LOCATION: 4 <223> OTHER INFORMATION: This
nucleotide may be absent <400> SEQUENCE: 132 cagatctcca ccac
14 <210> SEQ ID NO 133 <211> LENGTH: 19 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(9)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 133 gaagacagaa gttctgggt 19 <210> SEQ
ID NO 134 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 134 gggtctacca
ctgaa 15 <210> SEQ ID NO 135 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 135 gtgacagatg agagaa 16 <210>
SEQ ID NO 136 <211> LENGTH: 19 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(9) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 136 catacacatt caaacactt 19 <210> SEQ ID NO 137
<211> LENGTH: 19 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(9) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
137 acacattcaa acacttaca 19 <210> SEQ ID NO 138 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 138 cattcaaaca ctta 14 <210> SEQ
ID NO 139 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 139 cattcaaaca
cttac 15 <210> SEQ ID NO 140 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 140 aacacttaca
atttcac 17 <210> SEQ ID NO 141 <211> LENGTH: 22
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(12) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 141 tacaatttca
ctaagtcgga aa 22 <210> SEQ ID NO 142 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 142 ttcactaagt
cggaaaat 18 <210> SEQ ID NO 143 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 143 aagtcggaaa
attcaaa 17 <210> SEQ ID NO 144 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 144 acatgttcta tgcct 15 <210>
SEQ ID NO 145 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 145 ttagaataca agagat 16 <210> SEQ ID NO 146
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 146 aatgatagtt
taaa 14 <210> SEQ ID NO 147 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 147 agtttaaata gtgtca 16 <210>
SEQ ID NO 148 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 148 ttaaatagtg tcagtag 17 <210> SEQ ID NO 149
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 149 tatggtaaaa
gaggt 15 <210> SEQ ID NO 150 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: 6
<223> OTHER INFORMATION: This nucleotide may be absent
<400> SEQUENCE: 150 ggtaaaagag gtcaaa 16 <210> SEQ ID
NO 151 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(6) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
151 aaaagaggtc aaatga 16 <210> SEQ ID NO 152 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 152 agtaagtttt gcagtt 16 <210>
SEQ ID NO 153 <211> LENGTH: 24 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(14) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 153 aagttttgca gttatggtca atac 24 <210> SEQ ID NO
154 <211> LENGTH: 20 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(10) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
154 caatacccag ccgacctagc 20 <210> SEQ ID NO 155 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 155 acaccaataa attat 15 <210>
SEQ ID NO 156 <211> LENGTH: 14 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 4 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
156 aaatattcag atga 14 <210> SEQ ID NO 157 <211>
LENGTH: 19 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 157 tcagatgagc
agttgaact 19 <210> SEQ ID NO 158 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 158 gatgagcagt tgaac 15 <210>
SEQ ID NO 159 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 5 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
159 tgggcaagac ccaaa 15 <210> SEQ ID NO 160 <211>
LENGTH: 19 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 160 cacataatag
aagatgaaa 19 <210> SEQ ID NO 161 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 161 ataatagaag
atgaaataa 19 <210> SEQ ID NO 162 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 162 atagaagatg aaataa 16 <210>
SEQ ID NO 163 <211> LENGTH: 20 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(10) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 163 ataaaacaaa gtgagcaaag 20 <210> SEQ ID NO 164
<211> LENGTH: 17 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
164 aaacaaagtg agcaaag 17 <210> SEQ ID NO 165 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 165 aaacaaagtg
agcaaa 16 <210> SEQ ID NO 166 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 166 caaagtgagc
aaagacaa 18 <210> SEQ ID NO 167 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 167 caaagacaat
caaggaa 17 <210> SEQ ID NO 168 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 168 caatcaagga
atcaaag 17 <210> SEQ ID NO 169 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 169 caaagtacaa
cttatc 16 <210> SEQ ID NO 170 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 170 actgagagca
ctgatg 16 <210> SEQ ID NO 171 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 171 actgatgata
aacacct 17 <210> SEQ ID NO 172 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 172 gataaacacc
tcaagtt 17 <210> SEQ ID NO 173 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 173 cacctcaagt
tccaac 16 <210> SEQ ID NO 174 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 174 tttggacagc aggaa 15 <210>
SEQ ID NO 175 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(5) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 175 tgtgtttctc catac 15 <210> SEQ ID NO 176
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 176 cggggagcca
atgg 14 <210> SEQ ID NO 177 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 177 tcagaaacaa atcgag 16 <210>
SEQ ID NO 178 <211> LENGTH: 19 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(9) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 178 attaatcaaa atgtaagcc 19 <210> SEQ ID NO 179
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 179 caagaagatg
acta 14 <210> SEQ ID NO 180 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 180 gactatgaag atgata 16 <210>
SEQ ID NO 181 <211> LENGTH: 18 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(8) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 181 gatgataagc ctaccaat 18 <210> SEQ ID NO 182
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 182 cgttactctg
aagaag 16 <210> SEQ ID NO 183 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 183 gaagaagaag
agagaccaa 19 <210> SEQ ID NO 184 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 184 gaagaagaga
gaccaaca 18 <210> SEQ ID NO 185 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 185 gaagagagac
caacaaa 17 <210> SEQ ID NO 186 <211> LENGTH: 18
<212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(8) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 186 gaagagaaac gtcatgtg 18 <210> SEQ ID NO 187
<211> LENGTH: 22 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(12) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
187 gattatagtt taaaatatgc ca 22 <210> SEQ ID NO 188
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 188 ttaaaatatg
ccaca 15 <210> SEQ ID NO 189 <211> LENGTH: 18
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(8) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 189 gccacagata
ttccttca 18 <210> SEQ ID NO 190 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(5) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 190 acagatattc
cttca 15 <210> SEQ ID NO 191 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: These nucleotides may be
absent <400> SEQUENCE: 191 tcacagaaac agtcat 16 <210>
SEQ ID NO 192 <211> LENGTH: 15 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(5) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 192 aaacagtcat tttca 15 <210> SEQ ID NO 193
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
193 tcaaagagtt catct 15 <210> SEQ ID NO 194 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 194 aaaaccgaac atatg 15 <210>
SEQ ID NO 195 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 195 accgaacata tgtcttc 17 <210> SEQ ID NO 196
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be present <400> SEQUENCE:
196 catatgtctt caagc 15 <210> SEQ ID NO 197 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
present <400> SEQUENCE: 197 ccaagttctg cacaga 16 <210>
SEQ ID NO 198 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be present <400>
SEQUENCE: 198 tgcaaagttt cttcta 16 <210> SEQ ID NO 199
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be present <400> SEQUENCE:
199 atacagactt attgt 15 <210> SEQ ID NO 200 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(7) <223> OTHER INFORMATION: These
nucleotides may be present <400> SEQUENCE: 200 cagacttatt
gtgtaga 17 <210> SEQ ID NO 201 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 201 ccaatatgtt tttc 14 <210> SEQ
ID NO 202 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 202 agttcattat
catc 14 <210> SEQ ID NO 203 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 203 caggaagcag attctg 16 <210>
SEQ ID NO 204 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
present <400> SEQUENCE: 204 accctgcaaa tagca 15 <210>
SEQ ID NO 205 <211> LENGTH: 18 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(8) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 205 gaaataaaag aaaagatt 18 <210> SEQ ID NO 206
<211> LENGTH: 14 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 206 ataaaagaaa
agat 14 <210> SEQ ID NO 207 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(7) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 207 aaagaaaaga
ttggaac 17 <210> SEQ ID NO 208 <211> LENGTH: 20
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(10) <223> OTHER INFORMATION: These
nucleotides may be present <400> SEQUENCE: 208 aaagaaaaga
ttggaactag 20 <210> SEQ ID NO 209 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 209 gatcctgtga gcgaa 15 <210>
SEQ ID NO 210 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
210 gtgagcgaag ttccag 16 <210> SEQ ID NO 211 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
present <400> SEQUENCE: 211 gttccagcag tgtca 15 <210>
SEQ ID NO 212 <211> LENGTH: 22 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(13) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 212 caccctagaa ccaaatccag ca 22 <210> SEQ ID NO 213
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(6) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
213 agactgcagg gttcta 16 <210> SEQ ID NO 214 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 214 cagggttcta gttt 14 <210> SEQ
ID NO 215 <211> LENGTH: 15 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 215 tctagtttat
cttca 15 <210> SEQ ID NO 216 <211> LENGTH: 15
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 216 ttatcttcag aatca 15 <210>
SEQ ID NO 217 <211> LENGTH: 14 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 4 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
217 gttgaatttt cttc 14 <210> SEQ ID NO 218 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 218 ccctccaaaa gtggt 15 <210>
SEQ ID NO 219 <211> LENGTH: 17 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (4)...(7) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 219 agtggtgctc agacacc 17 <210> SEQ ID NO 220
<211> LENGTH: 17 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
220 agtccacctg aacacta 17 <210> SEQ ID NO 221 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 221 ccacctgaac actatg 16 <210>
SEQ ID NO 222
<211> LENGTH: 16 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(6) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
222 tatgttcagg agaccc 16 <210> SEQ ID NO 223 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 223 gatagttttg
agagtc 16 <210> SEQ ID NO 224 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 224 attgccagct ccgttc 16 <210>
SEQ ID NO 225 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
225 agtggcatta taagcc 16 <210> SEQ ID NO 226 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 226 agccctggac aaacc 15 <210>
SEQ ID NO 227 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
227 cctggacaaa ccatgc 16 <210> SEQ ID NO 228 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 228 atgccaccaa gcaga 15 <210>
SEQ ID NO 229 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
229 aaaaataaag caccta 16 <210> SEQ ID NO 230 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 230 gaaaagagag agag 14 <210> SEQ
ID NO 231 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
231 agagagagtg gacctaa 17 <210> SEQ ID NO 232 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 232 gagagtggac ctaag 15 <210>
SEQ ID NO 233 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (5)...(6) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 233 gagagtggac ctaagc 16 <210> SEQ ID NO 234
<211> LENGTH: 15 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(5) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
234 gagagtggac ctaag 15 <210> SEQ ID NO 235 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 235 gccacggaaa gtac 14 <210> SEQ
ID NO 236 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 236 acggaaagta
ctccag 16 <210> SEQ ID NO 237 <211> LENGTH: 14
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 237 ccagatggat tttc 14 <210> SEQ
ID NO 238 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (4)...(7) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
238 tcatccagcc tgagtgc 17 <210> SEQ ID NO 239 <211>
LENGTH: 20 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (4)...(10) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 239 ttaagaataa
tgcctccagt 20
<210> SEQ ID NO 240 <211> LENGTH: 17 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: (6)...(7)
<223> OTHER INFORMATION: These nucleotides may be absent
<400> SEQUENCE: 240 gaaacagaat cagagca 17 <210> SEQ ID
NO 241 <211> LENGTH: 20 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(10) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
241 tcaaatgaaa accaagagaa 20 <210> SEQ ID NO 242 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 242 gaaaaccaag agaa 14 <210> SEQ
ID NO 243 <211> LENGTH: 18 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (5)...(8) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
243 gagaaagagg cagaaaaa 18 <210> SEQ ID NO 244 <211>
LENGTH: 16 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (5)...(6) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 244 gaatgtatta
tttctg 16 <210> SEQ ID NO 245 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 245 ccagcccaga ctgctt 16 <210>
SEQ ID NO 246 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
246 cagactgctt caaaat 16 <210> SEQ ID NO 247 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 247 ttcaatgata agctc 15 <210>
SEQ ID NO 248 <211> LENGTH: 19 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: (6)...(9) <223>
OTHER INFORMATION: These nucleotides may be absent <400>
SEQUENCE: 248 aatgattctt tgagttctc 19 <210> SEQ ID NO 249
<211> LENGTH: 19 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: (6)...(9) <223> OTHER
INFORMATION: These nucleotides may be absent <400> SEQUENCE:
249 ccagacagag gggcagcaa 19 <210> SEQ ID NO 250 <211>
LENGTH: 14 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 4 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 250 gaaaatactc cagt 14 <210> SEQ
ID NO 251 <211> LENGTH: 14 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 4 <223> OTHER INFORMATION:
This nucleotide may be absent <400> SEQUENCE: 251 aacaataaag
aaaa 14 <210> SEQ ID NO 252 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 6 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 252 gaacctatca aagaga 16 <210>
SEQ ID NO 253 <211> LENGTH: 14 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 4 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
253 cctatcaaag agac 14 <210> SEQ ID NO 254 <211>
LENGTH: 15 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: This nucleotide may be
absent <400> SEQUENCE: 254 gaaccaagta aacct 15 <210>
SEQ ID NO 255 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Homo sapiens <220> FEATURE: <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: This nucleotide may be absent <400> SEQUENCE:
255 agctccgcaa tgccaa 16 <210> SEQ ID NO 256 <211>
LENGTH: 23 <212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(13) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 256 tcatcccttc
ctcgagtaag cac 23 <210> SEQ ID NO 257 <211> LENGTH: 19
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: (6)...(9) <223> OTHER INFORMATION: These
nucleotides may be absent <400> SEQUENCE: 257 ctaatttatc
aaatggcac 19
<210> SEQ ID NO 258 <211> LENGTH: 18 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: 6
<223> OTHER INFORMATION: n = C or G <221> NAME/KEY:
misc_feature <222> LOCATION: 7 <223> OTHER INFORMATION:
n = A or n is absent <221> NAME/KEY: misc_feature <222>
LOCATION: 8 <223> OTHER INFORMATION: n = A or n is absent
<400> SEQUENCE: 258 gaagannntt acagcagg 18 <210> SEQ ID
NO 259 <211> LENGTH: 18 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <220> FEATURE: <221> NAME/KEY:
misc_feature <222> LOCATION: 6 <223> OTHER INFORMATION:
n = T or C <221> NAME/KEY: misc_feature <222> LOCATION:
7 <223> OTHER INFORMATION: n = T <221> NAME/KEY:
misc_feature <222> LOCATION: 8 <223> OTHER INFORMATION:
n = A or n is absent <400> SEQUENCE: 259 cttacnnncc ggagccag
18 <210> SEQ ID NO 260 <211> LENGTH: 25 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <220> FEATURE:
<221> NAME/KEY: misc_feature <222> LOCATION: 4
<223> OTHER INFORMATION: n = C or T <221> NAME/KEY:
misc_feature <222> LOCATION: 5 <223> OTHER INFORMATION:
n = T <221> NAME/KEY: misc_feature <222> LOCATION: 6
<223> OTHER INFORMATION: n = C or G <221> NAME/KEY:
misc_feature <222> LOCATION: 7 <223> OTHER INFORMATION:
n = A or C <221> NAME/KEY: misc_feature <222> LOCATION:
8 <223> OTHER INFORMATION: n = T or A <221> NAME/KEY:
misc_feature <222> LOCATION: 9 <223> OTHER INFORMATION:
n = G or n is absent <221> NAME/KEY: misc_feature <222>
LOCATION: (10)...(10) <223> OTHER INFORMATION: n = C or n is
absent <221> NAME/KEY: misc_feature <222> LOCATION:
(11)...(13) <223> OTHER INFORMATION: n = T or n is absent
<221> NAME/KEY: misc_feature <222> LOCATION:
(14)...(15) <223> OTHER INFORMATION: n = A or n is absent
<400> SEQUENCE: 260 aatnnnnnnn nnnnnggcaa atagg 25
<210> SEQ ID NO 261 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <400> SEQUENCE: 261
ttgcagcttt aa 12 <210> SEQ ID NO 262 <211> LENGTH: 17
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<220> FEATURE: <221> NAME/KEY: misc_feature <222>
LOCATION: 5 <223> OTHER INFORMATION: n = G or T <221>
NAME/KEY: misc_feature <222> LOCATION: 6 <223> OTHER
INFORMATION: n = T or A <221> NAME/KEY: misc_feature
<222> LOCATION: 7 <223> OTHER INFORMATION: n = A or n
is absent <400> SEQUENCE: 262 gatgnnntat ggtaaaa 17
<210> SEQ ID NO 263 <211> LENGTH: 21 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 263
tggcgattaa gtcaaattcg c 21 <210> SEQ ID NO 264 <211>
LENGTH: 25 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 264 ccccctagta ccctgacaat gtatt 25
<210> SEQ ID NO 265 <211> LENGTH: 24 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 265
ctgttctgtg atattatgtg tggt 24 <210> SEQ ID NO 266 <211>
LENGTH: 22 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 266 aattgttggc attccagcat tg 22 <210>
SEQ ID NO 267 <211> LENGTH: 38 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 267
gtgcacttac gtgaattcag atgaacgtga tgtagtag 38 <210> SEQ ID NO
268 <211> LENGTH: 28 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 268 tcctcgtact
caacggcttt ctctgaat 28 <210> SEQ ID NO 269 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 269 tccggaacac tagaattctt atttacatac acacttgt
38 <210> SEQ ID NO 270 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 270 cgaataaggt agacggcaac aatgagaa 28 <210> SEQ ID
NO 271 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 271 cggtaaatcg
gagaattcag aggatttaga ggagctaa 38 <210> SEQ ID NO 272
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 272 ctcacgttcg ttacggccat
tgtgatagc 29 <210> SEQ ID NO 273 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 273 ggggaaacag tagaattcca tatggacaga gctgtact
38 <210> SEQ ID NO 274 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 274 tgaagctgtc ggacggcctt tgccctctc 29
<210> SEQ ID NO 275 <211> LENGTH: 38 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 275
atgggcagtt atgaattcac tactccctgt agcttgtt 38 <210> SEQ ID NO
276 <211> LENGTH: 29 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 276 tgattggcgc
gaacggcact cagagaaga 29 <210> SEQ ID NO 277 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 277 ctcaagggga ccgaattcgc tggggtcttc tgtgggtc
38 <210> SEQ ID NO 278 <211> LENGTH: 28 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 278 tagggcggcg tgacggccag ccagtggt 28 <210> SEQ ID
NO 279 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 279 gtcttgcatg
tagaattcta gggacgctgc ttttcgtc 38 <210> SEQ ID NO 280
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 280 ctcctagaca tcgggactag
aatgtccac 29 <210> SEQ ID NO 281 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 281 acacaaggca gagaattcca gtcctgaggg tgggggcc
38 <210> SEQ ID NO 282 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 282 ccgtgtttta acgggacaag ctgttcttc 29 <210> SEQ ID
NO 283 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 283 gtagcggagg
ttgaattcta tatgttgtct tggacatt 38 <210> SEQ ID NO 284
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 284 catcagtaga gtgggacgaa
agttctggc 29 <210> SEQ ID NO 285 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 285 atccacgccg cagaattcgt attcatgggc atgtcaaa
38 <210> SEQ ID NO 286 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 286 cttgggacta ttgggaccag tgttcaatc 29 <210> SEQ ID
NO 287 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 287 ccagaaagcc
gtgaattcgt taagccaacc tgactcca 38 <210> SEQ ID NO 288
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 288 tcggggttag tcgggacatc
cagcagccc 29 <210> SEQ ID NO 289 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 289 cgaaggtaat gtgaattcca aaacttagtg ccacaatt
38 <210> SEQ ID NO 290 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 290 ataccgccca acgggacaga tccattgac 29 <210> SEQ ID
NO 291 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 291 agaaacctgt
aagaattcga ttccaaattg ttttttgg 38 <210> SEQ ID NO 292
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 292 cgatcatagg gggggacagg
agagagcac 29 <210> SEQ ID NO 293 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 293 ctactgaggg ctcgtagatc ccaattcctt cccaagct
38 <210> SEQ ID NO 294 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 294 aatcctgctt tagggaccat gctggtgga 29 <210> SEQ ID
NO 295 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 295
caagtgtcct aa 12 <210> SEQ ID NO 296 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 296 cagctgctag aa 12 <210> SEQ ID NO
297 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 297 ggttgagggc aa
12 <210> SEQ ID NO 298 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 298 cacagcgggt aa 12 <210> SEQ ID NO 299
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 299 ttgacttttt aa 12
<210> SEQ ID NO 300 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 300
acagaatggg aa 12 <210> SEQ ID NO 301 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 301 tgcaggtcac aa 12 <210> SEQ ID NO
302 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 302 ttcttcttat aa
12 <210> SEQ ID NO 303 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 303 aggacaacct aa 12 <210> SEQ ID NO 304
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 304 tggtgttcag aa 12
<210> SEQ ID NO 305 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 305
tcagcatatg aa 12 <210> SEQ ID NO 306 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 306 gttgccacac aa 12 <210> SEQ ID NO
307 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 307 cccagctagc aa
12 <210> SEQ ID NO 308 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 308 gggtcactgt aa 12 <210> SEQ ID NO 309
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 309 ttaaataccc aa 12
<210> SEQ ID NO 310 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 310
ttaggaggtt aa 12 <210> SEQ ID NO 311 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 311 acacagaatc aa 12 <210> SEQ ID NO
312 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 312 cgctgaggtc aa
12 <210> SEQ ID NO 313 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 313 aagtagagtc aa 12 <210> SEQ ID NO 314
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 314 cttcccatgg aa 12
<210> SEQ ID NO 315 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 315
ttggttatta aa 12 <210> SEQ ID NO 316 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 316
caacttactg aa 12 <210> SEQ ID NO 317 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 317 cactaagtga aa 12 <210> SEQ ID NO
318 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 318 ctcacctgcc aa
12 <210> SEQ ID NO 319 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 319 atgcatatat aa 12 <210> SEQ ID NO 320
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 320 agagatcagc aa 12
<210> SEQ ID NO 321 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 321
tatatttttc aa 12 <210> SEQ ID NO 322 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 322 cagaaagcag aa 12 <210> SEQ ID NO
323 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 323 gtattgggtt aa
12 <210> SEQ ID NO 324 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 324 ctgacccagg aa 12 <210> SEQ ID NO 325
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 325 cagttttccc aa 12
<210> SEQ ID NO 326 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 326
agggcacagg aa 12 <210> SEQ ID NO 327 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 327 gtatcagagg aa 12 <210> SEQ ID NO
328 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 328 gcatgaaaag aa
12 <210> SEQ ID NO 329 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 329 gatttgacag aa 12 <210> SEQ ID NO 330
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 330 tacagtttac aa 12
<210> SEQ ID NO 331 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 331
tgtgattttt aa 12 <210> SEQ ID NO 332 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 332 ttatgttctc aa 12 <210> SEQ ID NO
333 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 333 caagtacttg aa
12 <210> SEQ ID NO 334 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 334 cttgtgtggc aa 12 <210> SEQ ID NO 335
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 335 agacttctgc aa 12
<210> SEQ ID NO 336 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 336
gttgtctttc aa 12 <210> SEQ ID NO 337 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 337 gggacactcc aa 12 <210> SEQ ID NO
338 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 338 attattattc aa
12 <210> SEQ ID NO 339 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 339 acatgatgac aa 12 <210> SEQ ID NO 340
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 340 tcaattatag aa 12
<210> SEQ ID NO 341 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 341
ctatgggctg aa 12 <210> SEQ ID NO 342 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 342 tgtgtgcctg aa 12 <210> SEQ ID NO
343 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 343 ccatttgttg aa
12 <210> SEQ ID NO 344 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 344 tctccatcaa aa 12 <210> SEQ ID NO 345
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 345 aatgctgaca aa 12
<210> SEQ ID NO 346 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 346
tttcatgtcc aa 12 <210> SEQ ID NO 347 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 347 ggcctcttgg aa 12 <210> SEQ ID NO
348 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 348 tcattttttg aa
12 <210> SEQ ID NO 349 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 349 ggactaccat aa 12 <210> SEQ ID NO 350
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 350 agtcactcag aa 12
<210> SEQ ID NO 351 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 351
ccttggcagg aa 12 <210> SEQ ID NO 352 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 352 tttctggtag aa 12 <210> SEQ ID NO
353 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 353 cccccccccg aa
12 <210> SEQ ID NO 354 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 354 gcccaggcag aa 12 <210> SEQ ID NO 355
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 355 gaatgcgaag aa 12
<210> SEQ ID NO 356 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 356
ttaggtagag aa 12 <210> SEQ ID NO 357 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 357 tgctttggtc aa 12 <210> SEQ ID NO
358 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer
<400> SEQUENCE: 358 gcccattaat aa 12 <210> SEQ ID NO
359 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 359 tgagatcttt aa
12 <210> SEQ ID NO 360 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 360 cagtttgttc aa 12 <210> SEQ ID NO 361
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 361 gctgggcaag aa 12
<210> SEQ ID NO 362 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 362
agtcaaagtc aa 12 <210> SEQ ID NO 363 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 363 tctctgcagt aa 12 <210> SEQ ID NO
364 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 364 tgaataactt aa
12 <210> SEQ ID NO 365 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 365 cggttagaaa aa 12 <210> SEQ ID NO 366
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 366 catccctttc aa 12
<210> SEQ ID NO 367 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 367
tctctttctg aa 12 <210> SEQ ID NO 368 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 368 ctcagattgt aa 12 <210> SEQ ID NO
369 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 369 tttgcaccag aa
12 <210> SEQ ID NO 370 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 370 ggttaacatg aa 12 <210> SEQ ID NO 371
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 371 attatcaact aa 12
<210> SEQ ID NO 372 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 372
gccattttgt aa 12 <210> SEQ ID NO 373 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 373 gatctagatg aa 12 <210> SEQ ID NO
374 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 374 ttaatgtatt aa
12 <210> SEQ ID NO 375 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 375 ctagggagac aa 12 <210> SEQ ID NO 376
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 376 tggaggagac aa 12
<210> SEQ ID NO 377 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 377
catcacattt aa 12 <210> SEQ ID NO 378 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 378 ggggtcctgc aa 12 <210> SEQ ID NO
379 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 379
cagttgtgct aa 12 <210> SEQ ID NO 380 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 380 tctgcagcct aa 12 <210> SEQ ID NO
381 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 381 gagtcattta aa
12 <210> SEQ ID NO 382 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 382 tctatggatt aa 12 <210> SEQ ID NO 383
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 383 caaaaagtag aa 12
<210> SEQ ID NO 384 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 384
aatatactcc aa 12 <210> SEQ ID NO 385 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 385 cgtccagcac aa 12 <210> SEQ ID NO
386 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 386 ggatggtgag aa
12 <210> SEQ ID NO 387 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 387 tctcctttgt aa 12 <210> SEQ ID NO 388
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 388 tcgttatttc aa 12
<210> SEQ ID NO 389 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 389
gattttatag aa 12 <210> SEQ ID NO 390 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 390 agacataagc aa 12 <210> SEQ ID NO
391 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 391 ttcacctcac aa
12 <210> SEQ ID NO 392 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 392 ggattgcttg aa 12 <210> SEQ ID NO 393
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 393 actgcatgtg aa 12
<210> SEQ ID NO 394 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 394
tttatcacag aa 12 <210> SEQ ID NO 395 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 395 tcagtaacac aa 12 <210> SEQ ID NO
396 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 396 tacatctttg aa
12 <210> SEQ ID NO 397 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 397 ttgtttcagt aa 12 <210> SEQ ID NO 398
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 398 tatgagcatc aa 12
<210> SEQ ID NO 399 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 399
ctcagcaggc aa 12 <210> SEQ ID NO 400 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 400 acccctgtat aa 12 <210> SEQ ID NO
401 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 401 tctgctcagc aa
12 <210> SEQ ID NO 402 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 402 gttctttttt aa 12 <210> SEQ ID NO 403
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 403 gtgataatcc aa 12
<210> SEQ ID NO 404 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 404
gagccctcag aa 12 <210> SEQ ID NO 405 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 405 tttattggtt aa 12 <210> SEQ ID NO
406 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 406 ggtactgggc aa
12 <210> SEQ ID NO 407 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 407 agtgtttttc aa 12 <210> SEQ ID NO 408
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 408 tgttattggt aa 12
<210> SEQ ID NO 409 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 409
gcgcattcac aa 12 <210> SEQ ID NO 410 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 410 aaacaaaagc aa 12 <210> SEQ ID NO
411 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 411 tatatgatag aa
12 <210> SEQ ID NO 412 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 412 tcccagttcc aa 12 <210> SEQ ID NO 413
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 413 aaagcccata aa 12
<210> SEQ ID NO 414 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 414
tgtcatccac aa 12 <210> SEQ ID NO 415 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 415 ttgtgaatgc aa 12 <210> SEQ ID NO
416 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 416 gtattcatac aa
12 <210> SEQ ID NO 417 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 417 tgacataggg aa 12 <210> SEQ ID NO 418
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 418 agcaaattgc aa 12
<210> SEQ ID NO 419 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 419
agtagatgtt aa 12 <210> SEQ ID NO 420 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 420 aaaagataat aa 12 <210> SEQ ID NO
421 <211> LENGTH: 12 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 421
acctcatggg aa 12 <210> SEQ ID NO 422 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 422 tggtcgacct aa 12 <210> SEQ ID NO
423 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 423 tttgcatggt aa
12 <210> SEQ ID NO 424 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 424 gcggctgccg aa 12 <210> SEQ ID NO 425
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 425 tcaggagtct aa 12
<210> SEQ ID NO 426 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 426
gcctaccagg aa 12 <210> SEQ ID NO 427 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 427 atcttctgtt aa 12 <210> SEQ ID NO
428 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 428 aggtaaggac aa
12 <210> SEQ ID NO 429 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 429 tgctttgagg aa 12 <210> SEQ ID NO 430
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 430 aacagtttta aa 12
<210> SEQ ID NO 431 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 431
ttaaatgttt aa 12 <210> SEQ ID NO 432 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 432 atagaaaatc aa 12 <210> SEQ ID NO
433 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 433 gtgttgtgtt aa
12 <210> SEQ ID NO 434 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 434 gaggacctcg aa 12 <210> SEQ ID NO 435
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 435 agaggctgag aa 12
<210> SEQ ID NO 436 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 436
ggtatttatt aa 12 <210> SEQ ID NO 437 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 437 atttatctgg aa 12 <210> SEQ ID NO
438 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 438 agtgcaaact aa
12 <210> SEQ ID NO 439 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 439 tgaacacctt aa 12 <210> SEQ ID NO 440
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 440 aattttttct aa 12
<210> SEQ ID NO 441 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 441
ttactattat aa 12 <210> SEQ ID NO 442 <211> LENGTH:
12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 442 tgctatagtg aa 12 <210> SEQ ID NO
443 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 443 tggactatgg aa
12 <210> SEQ ID NO 444 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 444 ctgcagtccg aa 12 <210> SEQ ID NO 445
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 445 gctactgccc aa 12
<210> SEQ ID NO 446 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 446
tcacatggtg aa 12 <210> SEQ ID NO 447 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 447 gtggctctgg aa 12 <210> SEQ ID NO
448 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 448 gaattccatt aa
12 <210> SEQ ID NO 449 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 449 tggggtgtcc aa 12 <210> SEQ ID NO 450
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 450 gcaagctccg aa 12
<210> SEQ ID NO 451 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 451
atgttttttc aa 12 <210> SEQ ID NO 452 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 452 agatctgttg aa 12 <210> SEQ ID NO
453 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 453 aagtgctgtg aa
12 <210> SEQ ID NO 454 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 454 acttttttgg aa 12 <210> SEQ ID NO 455
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 455 aatcggcagg aa 12
<210> SEQ ID NO 456 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 456
ggcatgtcac aa 12 <210> SEQ ID NO 457 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 457 aggaagaaag aa 12 <210> SEQ ID NO
458 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 458 cagtttcacc aa
12 <210> SEQ ID NO 459 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 459 cacagaattt aa 12 <210> SEQ ID NO 460
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 460 aagaataagt aa 12
<210> SEQ ID NO 461 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 461
gggatagtac aa 12 <210> SEQ ID NO 462 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 462 ttcccatgat aa 12 <210> SEQ ID NO
463
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 463 tgattagttg aa 12
<210> SEQ ID NO 464 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 464
gcattcagtg aa 12 <210> SEQ ID NO 465 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 465 agggaatatt aa 12 <210> SEQ ID NO
466 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 466 gaccttaggt aa
12 <210> SEQ ID NO 467 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 467 ttcttttcac aa 12 <210> SEQ ID NO 468
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 468 ccaaactaag aa 12
<210> SEQ ID NO 469 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 469
gtgctcttag aa 12 <210> SEQ ID NO 470 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 470 atgagtttag aa 12 <210> SEQ ID NO
471 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 471 atgagcatag aa
12 <210> SEQ ID NO 472 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 472 gacaaatgag aa 12 <210> SEQ ID NO 473
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 473 aaacccagag aa 12
<210> SEQ ID NO 474 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 474
cctcacacag aa 12 <210> SEQ ID NO 475 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 475 cacactgtgg aa 12 <210> SEQ ID NO
476 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 476 cactgtaccc aa
12 <210> SEQ ID NO 477 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 477 gtagtatttc aa 12 <210> SEQ ID NO 478
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 478 tggatacact aa 12
<210> SEQ ID NO 479 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 479
cccatgattc aa 12 <210> SEQ ID NO 480 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 480 tcataggagg aa 12 <210> SEQ ID NO
481 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 481 aggaaagaga aa
12 <210> SEQ ID NO 482 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 482 atatggtgat aa 12 <210> SEQ ID NO 483
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 483 gatgccatcc aa 12
<210> SEQ ID NO 484 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 484
atactatttc aa 12 <210> SEQ ID NO 485 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 485 gtgtgcatgg aa 12 <210> SEQ ID NO
486 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 486 aggtgttgag aa
12 <210> SEQ ID NO 487 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 487 cagcctgggc aa 12 <210> SEQ ID NO 488
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 488 ggagctctac aa 12
<210> SEQ ID NO 489 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 489
aactaaggtt aa 12 <210> SEQ ID NO 490 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 490 aacttatgtt aa 12 <210> SEQ ID NO
491 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 491 atctcaacag aa
12 <210> SEQ ID NO 492 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 492 taacaatgtg aa 12 <210> SEQ ID NO 493
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 493 aaggatcagg aa 12
<210> SEQ ID NO 494 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 494
ctcaagtctt aa 12 <210> SEQ ID NO 495 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 495 ttagtcatcg cagaattcta cttctttctg aagtggga
38 <210> SEQ ID NO 496 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 496 ggacagctcg atgggactaa tgcatactc 29 <210> SEQ ID
NO 497 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 497 gtagccactg
gtgaattcgt gccatcgcaa aagaataa 38 <210> SEQ ID NO 498
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 498 attagaatga tggggacccc
tgtcttccc 29 <210> SEQ ID NO 499 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 499 acgcatagga aggaattcat tctgacacgt gtgagata
38 <210> SEQ ID NO 500 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 500 gaaattgacc acgggactgc acacttttc 29 <210> SEQ ID
NO 501 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 501 cggtaaatcg
gagaattcaa gttgaggcat gcatccat 38 <210> SEQ ID NO 502
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 502 tcggggctca gcgggaccac
agccactcc 29 <210> SEQ ID NO 503 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 503 tctatgcacc acgaattcaa tatgtgttca aggacatt
38 <210> SEQ ID NO 504 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 504 tgcttaatcg gtgggacttg taattgtac 29
<210> SEQ ID NO 505 <211> LENGTH: 38 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 505
cgcgttgtat gcgaattccc tggggtataa agataaga 38 <210> SEQ ID NO
506 <211> LENGTH: 29 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 506 ctcacgggaa
ctgggacacc tgaccctgc 29 <210> SEQ ID NO 507 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 507 gtcttgccgc ttgaattccc atagaagaat gcgccaaa
38 <210> SEQ ID NO 508 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 508 ttgagtagta cagggacaca ctaacagac 29 <210> SEQ ID
NO 509 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 509 aatactgtag
gtgaattctt gcctaagcat tttcccag 38 <210> SEQ ID NO 510
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 510 gtgttgacat tcgggactgt
aatcttgac 29 <210> SEQ ID NO 511 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 511 tctgtagatt cggaattctt tagagcctgt gcgctgag
38 <210> SEQ ID NO 512 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 512 cgtaccagta cagggacgca aactgagac 29 <210> SEQ ID
NO 513 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 513 gacatgttgg aa
12 <210> SEQ ID NO 514 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 514 acttccagtt aa 12 <210> SEQ ID NO 515
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 515 gtttcctgtt aa 12
<210> SEQ ID NO 516 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 516
cgatgatgac aa 12 <210> SEQ ID NO 517 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 517 gagtagagac aa 12 <210> SEQ ID NO
518 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 518 tcccggatac aa
12 <210> SEQ ID NO 519 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 519 catcctctag aa 12 <210> SEQ ID NO 520
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 520 tattcctgag aa 12
<210> SEQ ID NO 521 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 521
agtttgtttt aa 12 <210> SEQ ID NO 522 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 522 tataaacgat aa 12 <210> SEQ ID NO
523 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 523 tttgaccgat aa
12 <210> SEQ ID NO 524 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 524 tgacaggacc aa 12 <210> SEQ ID NO 525
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 525 ttattcattc aa 12
<210> SEQ ID NO 526 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 526
agtttttcac aa 12 <210> SEQ ID NO 527 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 527 cacctccctg aa 12 <210> SEQ ID NO
528 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 528 ccagattgag aa
12 <210> SEQ ID NO 529 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 529 tgtgtccacc aa 12 <210> SEQ ID NO 530
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 530 cttctattcc aa 12
<210> SEQ ID NO 531 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 531
tcacaatagg aa 12 <210> SEQ ID NO 532 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 532 tacaagtgag aa 12 <210> SEQ ID NO
533 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 533 gagttttcgt aa
12 <210> SEQ ID NO 534 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 534 gtgtgccccc aa 12 <210> SEQ ID NO 535
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 535 gcaccactgc aa 12
<210> SEQ ID NO 536 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 536
gaacacaatg aa 12 <210> SEQ ID NO 537 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 537 tatcctattc aa 12 <210> SEQ ID NO
538 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 538 caaccattat aa
12 <210> SEQ ID NO 539 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 539 tatgctttac aa 12 <210> SEQ ID NO 540
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 540 tttgtttacc aa 12
<210> SEQ ID NO 541 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 541
aggaaattag aa 12 <210> SEQ ID NO 542 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 542 tgttagactt aa 12 <210> SEQ ID NO
543 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 543 tatttggagg aa
12 <210> SEQ ID NO 544 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 544 ggcatttgtc aa 12 <210> SEQ ID NO 545
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 545 atactccagg aa 12
<210> SEQ ID NO 546 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 546
cagcctggac aa 12 <210> SEQ ID NO 547 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 547 ccattgcagt aa 12 <210> SEQ ID NO
548 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 548 aggttctcat aa
12 <210> SEQ ID NO 549 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 549 tgtcatcatt aa 12 <210> SEQ ID NO 550
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 550 tggtatttgc aa 12
<210> SEQ ID NO 551 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 551
tagggtttgt aa 12 <210> SEQ ID NO 552 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 552 ccctaagtag aa 12 <210> SEQ ID NO
553 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 553 gtatttcttt aa
12 <210> SEQ ID NO 554 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 554 gagtcttccc aa 12 <210> SEQ ID NO 555
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 555 caggtagagt aa 12
<210> SEQ ID NO 556 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 556
ataggatgtg aa 12 <210> SEQ ID NO 557 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 557 caatgtgtat aa 12 <210> SEQ ID NO
558 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 558 agagggcatc aa
12 <210> SEQ ID NO 559 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 559 ccagtggtct aa 12 <210> SEQ ID NO 560
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 560 taaacaatag aa 12
<210> SEQ ID NO 561 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 561
gcacactttt aa 12 <210> SEQ ID NO 562 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 562 atggctctgc aa 12 <210> SEQ ID NO
563 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 563 gtcatcttgt aa
12 <210> SEQ ID NO 564 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 564 tgcttcatct aa 12 <210> SEQ ID NO 565
<211> LENGTH: 12 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 565 agaaaggggc aa 12
<210> SEQ ID NO 566 <211> LENGTH: 12 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 566
cttttctttc aa 12 <210> SEQ ID NO 567 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 567
ctactctctc aa 12 <210> SEQ ID NO 568 <211> LENGTH: 12
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 568 acagcattat aa 12 <210> SEQ ID NO
569 <211> LENGTH: 12 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 569 actgctctgg aa
12 <210> SEQ ID NO 570 <211> LENGTH: 12 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 570 gcagaggcac aa 12 <210> SEQ ID NO 571
<211> LENGTH: 38 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 571 ctccgtggta tggaattcca
ctcaaatctt cattcaga 38 <210> SEQ ID NO 572 <211>
LENGTH: 29 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 572 acgtcgggtt acgggacacc tgattcctc 29
<210> SEQ ID NO 573 <211> LENGTH: 38 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 573
taccattggt ttgaattctt gtttcctgtt aaccatgc 38 <210> SEQ ID NO
574 <211> LENGTH: 29 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 574 gccgagttct
acgggacaga aaagggagc 29 <210> SEQ ID NO 575 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 575 tgcagtgatt tcgaattcga gacaatgctg cccagtca
38 <210> SEQ ID NO 576 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 576 tctaaattct ctgggaccat tccttcaac 29 <210> SEQ ID
NO 577 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 577 actaacagca
ctgaattcca tgctcttgga ctttccat 38 <210> SEQ ID NO 578
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 578 tcccctaacg ttgggacaca
gaatactac 29 <210> SEQ ID NO 579 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 579 gtcgacgatg gcgaattcct gccactcatt cagttagc
38 <210> SEQ ID NO 580 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 580 gaacggccca cagggacctg gcataactc 29 <210> SEQ ID
NO 581 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 581 tcatggtagc
aggaattctg ctttgaccga taaggaga 38 <210> SEQ ID NO 582
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 582 actgtgggat tcgggactgt
ctactaccc 29 <210> SEQ ID NO 583 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 583 acctctcggc cggaattcgg aaaagtgtac agatcatt
38 <210> SEQ ID NO 584 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 584 gccggatacg aagggacggc tcgtgactc 29 <210> SEQ ID
NO 585 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 585 ccgtagacta
aagaattccc tgatgtcagg ctgtcacc 38 <210> SEQ ID NO 586
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 586 atcggatcag tcgggacggt
gtctttgcc 29 <210> SEQ ID NO 587 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 587 gcataggcgg gagaattccc tgtgtccacc aaagtcgg
38 <210> SEQ ID NO 588 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 588 cccacatagg gcgggacaaa gagctgaac 29
<210> SEQ ID NO 589 <211> LENGTH: 38 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 589
ggcttgccga gcgaattcta ggaaagatac ggaatcaa 38 <210> SEQ ID NO
590 <211> LENGTH: 29 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 590 taaccctcat
acgggacttt catggaagc 29 <210> SEQ ID NO 591 <211>
LENGTH: 38 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 591 atgagcaccc gggaattctg attggagtct aggccaaa
38 <210> SEQ ID NO 592 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 592 tgctcacctt ctgggacgtg gctggtctc 29 <210> SEQ ID
NO 593 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 593 accgtctgcc
acgaattctg gaaaacatgc agtctggt 38 <210> SEQ ID NO 594
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 594 tacacgggag gcgggacagg
gtgattaac 29 <210> SEQ ID NO 595 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 595 cttaaagcta acgaattcag agctgtatga agatgctt
38 <210> SEQ ID NO 596 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 596 aacgctaaag gggggacaac ataattggc 29 <210> SEQ ID
NO 597 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 597 ttgtaagaac
gagaattctg caacctgtct ttattgaa 38 <210> SEQ ID NO 598
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 598 cttcaccact ttgggacact
gaagccaac 29 <210> SEQ ID NO 599 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 599 aaccattgat ttgaattcga aatgtccacc aaagttca
38 <210> SEQ ID NO 600 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 600 tgtctagttc cagggacgct gttacttac 29 <210> SEQ ID
NO 601 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 601 cgaaggtaat
gtgaattctg ccacaattaa gacttgga 38 <210> SEQ ID NO 602
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 602 ataccggttt tcgggacaga
tccattgac 29 <210> SEQ ID NO 603 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 603 cctgaaatcc acgaattcca ccctggcctc ccagtgca
38 <210> SEQ ID NO 604 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 604 tagatggtag gtgggacagg actggcttc 29 <210> SEQ ID
NO 605 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 605 gcatatctta
gcgaattcct gtgactaata cagagtgc 38 <210> SEQ ID NO 606
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 606 ccaaatatgg tagggacgtg
tgaacactc 29 <210> SEQ ID NO 607 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 607 tgccgctaca gggaattcat atggcagata ttcctgaa
38 <210> SEQ ID NO 608 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 608 acgttgcgga ccgggacttc cacagagcc 29 <210> SEQ ID
NO 609 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer
<400> SEQUENCE: 609 cttcgcccaa tggaattcgg tacaggggta tgccttat
38 <210> SEQ ID NO 610 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 610 tgcacttctg ccgggaccag aggagaaac 29 <210> SEQ ID
NO 611 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 611 tgtgggtatt
ctgaattcca caaaatggac taacacgc 38 <210> SEQ ID NO 612
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 612 acgtcgttca gtgggacatt
aaaaggctc 29 <210> SEQ ID NO 613 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 613 ggttatgtgt cagaattctg aaactagttt ggaagtac
38 <210> SEQ ID NO 614 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 614 gcctcagttt cggggacagt tctgaggac 29 <210> SEQ ID
NO 615 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 615 tgtaacacgg
ccgaattcct catttgtatg aaataggt 38 <210> SEQ ID NO 616
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 616 aatctaactt gagggaccgg
cacacacac 29 <210> SEQ ID NO 617 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 617 agtgtcccct tagaattcgc agagacacca cagtgtgc
38 <210> SEQ ID NO 618 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 618 tttgctacag tcgggaccct tgtgtgctc 29 <210> SEQ ID
NO 619 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 619 agcacatcac
tagaattcaa taccatgtgt gagctcaa 38 <210> SEQ ID NO 620
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 620 aatcctgctt ccgggaccta
actttgaac 29 <210> SEQ ID NO 621 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 621 tttcattttc tggaattcct ctaatgattt tctggagc
38 <210> SEQ ID NO 622 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 622 cgtcgccgcg tagggacttt ttcttccac 29 <210> SEQ ID
NO 623 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 623 ttacttaatc
ctgaattcga gaaaagccat gttgataa 38 <210> SEQ ID NO 624
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 624 tcatgggtcg ctgggacttt
gccctctgc 29 <210> SEQ ID NO 625 <211> LENGTH: 38
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 625 actaacagca ctgaattcat tttactataa tctgctac
38 <210> SEQ ID NO 626 <211> LENGTH: 29 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Primer <400>
SEQUENCE: 626 gttagccgag aagggactgt ctgtgaagc 29 <210> SEQ ID
NO 627 <211> LENGTH: 38 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 627 aaatatgcag
cggaattcgt aagtgaccta ttaataac 38 <210> SEQ ID NO 628
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 628 gcgatggtta cggggacagc
caggcaacc 29 <210> SEQ ID NO 629 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 629 ccagctggta gaactt 16 <210> SEQ ID
NO 630 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 630
cccaatagac ctatag 16 <210> SEQ ID NO 631 <211> LENGTH:
16 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 631 tagcagaatc tctcat 16 <210> SEQ ID
NO 632 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 632 agagtatctc
atttgtt 17 <210> SEQ ID NO 633 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 633 aggaaattgt gaagta 16 <210> SEQ ID
NO 634 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 634 taactcactc
actatc 16 <210> SEQ ID NO 635 <211> LENGTH: 16
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 635 ctgctgagtc atagtc 16 <210> SEQ ID
NO 636 <211> LENGTH: 16 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 636 tgttctttga
atcaac 16 <210> SEQ ID NO 637 <211> LENGTH: 47
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 637 cagtaatacg actcactata ggggtcagga
ttagcagaat ctctcat 47 <210> SEQ ID NO 638 <211> LENGTH:
27 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 638 gcattctatg agagtatctc atttgtt 27
<210> SEQ ID NO 639 <211> LENGTH: 47 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 639
cagtaatacg actcactata ggggtcagga agagtatctc atttgtt 47 <210>
SEQ ID NO 640 <211> LENGTH: 26 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 640
gcattctatg tagcagaatc tctcat 26 <210> SEQ ID NO 641
<211> LENGTH: 17 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <221> NAME/KEY: misc_feature <222>
LOCATION: (1)...(17) <223> OTHER INFORMATION: n = T or A
<400> SEQUENCE: 641 ccgcatanct cagcaca 17 <210> SEQ ID
NO 642 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(17) <223> OTHER INFORMATION: n =
A or T <400> SEQUENCE: 642 tgtgctgagn tatgcgg 17 <210>
SEQ ID NO 643 <211> LENGTH: 16 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 643
aggaaattgt gaagta 16 <210> SEQ ID NO 644 <211> LENGTH:
16 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 644 taactcactc actatc 16 <210> SEQ ID
NO 645 <211> LENGTH: 25 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 645 ctttttatgc
ctttccactc atcca 25 <210> SEQ ID NO 646 <211> LENGTH:
25 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 646 ctttttatgc ctttccactc atcca 25
<210> SEQ ID NO 647 <211> LENGTH: 29 <212> TYPE:
DNA <213> ORGANISM: Homo sapiens <400> SEQUENCE: 647
gaaaaatacg gaaaggtgag taggtttcc 29 <210> SEQ ID NO 648
<211> LENGTH: 26 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 648 ctttttatgc ctttccactc
atccaa 26 <210> SEQ ID NO 649 <211> LENGTH: 29
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<400> SEQUENCE: 649 gaaaaatacg gaaaggtgag taggtttcc 29
<210> SEQ ID NO 650 <211> LENGTH: 25 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 650
ctttttatgc ctttccactc atcca 25 <210> SEQ ID NO 651
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Homo sapiens <400> SEQUENCE: 651
gaaaaatacg gaaaggtgag taggtctcc 29 <210> SEQ ID NO 652
<211> LENGTH: 27 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Primer <400> SEQUENCE: 652 ctttttatgc ctttccactc
atccaga 27 <210> SEQ ID NO 653 <211> LENGTH: 29
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<400> SEQUENCE: 653 gaaaaatacg gaaaggtgag taggtctcc 29
<210> SEQ ID NO 654 <211> LENGTH: 15 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 654
atcacactgg ggatc 15 <210> SEQ ID NO 655 <211> LENGTH:
15 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 655 ctaaacctat gactc 15 <210> SEQ ID NO
656 <211> LENGTH: 21 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: Primer <400> SEQUENCE: 656 ttcacagagg
atatcttaat a 21 <210> SEQ ID NO 657 <211> LENGTH: 22
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 657 agtcttgtaa tacgacagtc tt 22 <210>
SEQ ID NO 658 <211> LENGTH: 21 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 658
ccatatcaat cagtactctt g 21 <210> SEQ ID NO 659 <211>
LENGTH: 23 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION:
Oligonucleotide probe <400> SEQUENCE: 659 cctttccact
catccaaagg ttg 23 <210> SEQ ID NO 660 <211> LENGTH: 23
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Oligonucleotide
probe <400> SEQUENCE: 660 cctttccact catccagagg ttg 23
<210> SEQ ID NO 661 <211> LENGTH: 20 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Oligonucleotide probe <400>
SEQUENCE: 661 catgcagata taccgcatat 20 <210> SEQ ID NO 662
<211> LENGTH: 20 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Oligonucleotide probe <400> SEQUENCE: 662
catgcagata taccgcataa 20 <210> SEQ ID NO 663 <211>
LENGTH: 23 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION:
Oligonucleotide probe <400> SEQUENCE: 663 ggtagcatct
ctcagcacaa gag 23 <210> SEQ ID NO 664 <211> LENGTH: 23
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Oligonucleotide
probe <400> SEQUENCE: 664 ggtagcatca ctcagcacaa gag 23
<210> SEQ ID NO 665 <211> LENGTH: 16 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Primer <400> SEQUENCE: 665
tagcagaatc tctcat 16 <210> SEQ ID NO 666 <211> LENGTH:
17 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Primer
<400> SEQUENCE: 666 agagtatctc atttgtt 17 <210> SEQ ID
NO 667 <211> LENGTH: 125 <212> TYPE: DNA <213>
ORGANISM: Homo sapiens <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(125) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 667 gccaagtata attttcctga
taaaggctgg gctgcnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnncaattt gcactartga agtttaccta 120 acaat
125 <210> SEQ ID NO 668 <211> LENGTH: 125 <212>
TYPE: DNA <213> ORGANISM: Homo sapiens <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(125) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 668 attgttaggt
aaacttcayt agtgcaaatt gnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn gcagcccagc ctttatcagg aaaattatac
120 ttggc 125 <210> SEQ ID NO 669 <211> LENGTH: 140
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(140) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 669 ctatatgtga aagcagaagt tgtctgataa
tcnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnccaagg 120 sctttactcg
atgatagctg 140 <210> SEQ ID NO 670 <211> LENGTH: 140
<212> TYPE: DNA <213> ORGANISM: Homo sapiens
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(140) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 670 cagctatcat cgagtaaags ccttggnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnga ttatcagaca 120 acttctgctt
tcacatatag 140 <210> SEQ ID NO 671
<211> LENGTH: 141 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR Product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(141) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 671 ataaccgtat gcgaattcta
taattttcct gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn caatttgcac tartgaagtt 120
tacctagtcc ccagatttaa g 141 <210> SEQ ID NO 672 <211>
LENGTH: 141 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: PCR
Product <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(141) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 672 cttaaatcag gggactaggt aaacttcayt
agtgcaaatt gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn cagcctttat caggaaaatt 120 atagaattcg
catacggtta t 141 <210> SEQ ID NO 673 <211> LENGTH: 141
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(141) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 673 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tartgaagtt 120 tacctagccg
tcagatttaa g 141 <210> SEQ ID NO 674 <211> LENGTH: 141
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(141) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 674 cttaaatcag acggctaggt aaacttcayt
agtgcaaatt gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn cagcctttat caggaaaatt 120 atagaattcg
catacggtta t 141 <210> SEQ ID NO 675 <211> LENGTH: 161
<212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR Product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(161) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 675 aagtttagat cagaattcgt gaaagcagaa
gttgtctgat aatcnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 120 nnnnnnccaa
ggsctttact cgatgagtcc cttatcgtga t 161 <210> SEQ ID NO 676
<211> LENGTH: 161 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR Product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(161) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 676 atcacgataa gggactcatc
gagtaaagsc cttggnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnngat 120
tatcagacaa cttctgcttt cacgaattct gatctaaact t 161 <210> SEQ
ID NO 677 <211> LENGTH: 161 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR Product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(161) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 677 aagtttagat cagaattcgt
gaaagcagaa gttgtctgat aatcnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 120
nnnnnnccaa ggsctttact cgatgagccg tttatcgtga t 161 <210> SEQ
ID NO 678 <211> LENGTH: 161 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR Product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(161) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 678 atcacgataa acggctcatc
gagtaaagsc cttggnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnngat 120
tatcagacaa cttctgcttt cacgaattct gatctaaact t 161 <210> SEQ
ID NO 679 <211> LENGTH: 112 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(112) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 679 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac ta 112 <210> SEQ ID NO 680
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 680 rtgaagttta cctagtcccc agatttaag 29
<210> SEQ ID NO 681 <211> LENGTH: 116 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(116) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 681 tcaytagtgc aaattgnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnncagcc
tttatcagga aaattataga attcgcatac ggttat 116 <210> SEQ ID NO
682 <211> LENGTH: 25 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 682 cttaaatcag gggactaggt aaact 25
<210> SEQ ID NO 683 <211> LENGTH: 29 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 683 rtgaagttta cctagccgtc agatttaag
29 <210> SEQ ID NO 684 <211> LENGTH: 114 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: PCR product after
restriction enzyme digestion <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(114) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 684 aytagtgcaa attgnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnncagcctt tatcaggaaa attatagaat tcgcatacgg ttat 114 <210>
SEQ ID NO 685 <211> LENGTH: 27 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion
<400> SEQUENCE: 685 cttaaatcag acggctaggt aaacttc 27
<210> SEQ ID NO 686 <211> LENGTH: 132 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(132) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 686 aagtttagat cagaattcgt gaaagcagaa
gttgtctgat aatcnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 120 nnnnnnccaa gg 132
<210> SEQ ID NO 687 <211> LENGTH: 29 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 687 sctttactcg atgagtccct tatcgtgat
29 <210> SEQ ID NO 688 <211> LENGTH: 136 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: PCR product after
restriction enzyme digestion <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(136) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 688 aagsccttgg nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nngattatca gacaacttct gctttcacga 120
attctgatct aaactt 136 <210> SEQ ID NO 689 <211> LENGTH:
25 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: PCR product
after restriction enzyme digestion <400> SEQUENCE: 689
atcacgataa gggactcatc gagta 25 <210> SEQ ID NO 690
<211> LENGTH: 29 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 690 sctttactcg atgagccgtt tatcgtgat 29
<210> SEQ ID NO 691 <211> LENGTH: 134 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(134) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 691 gsccttggnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn gattatcaga caacttctgc tttcacgaat 120 tctgatctaa actt 134
<210> SEQ ID NO 692 <211> LENGTH: 27 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 692 atcacgataa acggctcatc gagtaaa
27 <210> SEQ ID NO 693 <211> LENGTH: 113 <212>
TYPE: DNA <213> ORGANISM: Artificial Sequence <220>
FEATURE: <223> OTHER INFORMATION: Polymerase extension
product <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(113) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 693 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tar 113 <210> SEQ ID NO 694
<211> LENGTH: 116 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Polymerase extension product <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(116) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 694 tcaytagtgc
aaattgnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnncagcc tttatcagga aaattataga attcgcatac ggttat 116
<210> SEQ ID NO 695 <211> LENGTH: 114 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Polymerase extension product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(114) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 695 aytagtgcaa attgnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnncagcctt
tatcaggaaa attatagaat tcgcatacgg ttat 114 <210> SEQ ID NO 696
<211> LENGTH: 133 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Polymerase extension product <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(133) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 696 aagtttagat
cagaattcgt gaaagcagaa gttgtctgat aatcnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn
120 nnnnnnccaa ggs 133 <210> SEQ ID NO 697 <211>
LENGTH: 136 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION:
Polymerase extension product <221> NAME/KEY: misc_feature
<222> LOCATION: (1)...(136) <223> OTHER INFORMATION: n
= A,T,C or G <400> SEQUENCE: 697 aagsccttgg nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nngattatca gacaacttct gctttcacga 120
attctgatct aaactt 136 <210> SEQ ID NO 698 <211> LENGTH:
134 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Polymerase
extension product <221> NAME/KEY: misc_feature <222>
LOCATION: (1)...(134) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 698 gsccttggnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn gattatcaga caacttctgc tttcacgaat 120 tctgatctaa actt 134
<210> SEQ ID NO 699 <211> LENGTH: 114 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: Polymerase extension product
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(114) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 699 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tart 114 <210> SEQ ID NO 700
<211> LENGTH: 115 <212> TYPE: DNA <213> ORGANISM:
Artificial Sequence <220> FEATURE: <223> OTHER
INFORMATION: Polymerase extension product <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(115)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
700 ataaccgtat gcgaattcta taattttcct gataaaggct gnnnnnnnnn
nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn
caatttgcac tartg 115 <210> SEQ ID NO 701 <211> LENGTH:
116 <212> TYPE: DNA <213> ORGANISM: Artificial Sequence
<220> FEATURE: <223> OTHER INFORMATION: Polymerase
extension product <221> NAME/KEY: misc_feature <222>
LOCATION: (1)...(116) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 701 ataaccgtat gcgaattcta taattttcct
gataaaggct gnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn caatttgcac tartga 116 <210> SEQ ID NO
702 <211> LENGTH: 13 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 702 ataaccgtat gcg 13 <210> SEQ ID NO
703 <211> LENGTH: 100 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(100) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 703 aattctataa ttttcctgat aaaggctgnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnncaa tttgcactar 100 <210> SEQ ID NO 704 <211>
LENGTH: 17 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: PCR
product after restriction enzyme digestion <400> SEQUENCE:
704 aattcgcata cggttat 17 <210> SEQ ID NO 705 <211>
LENGTH: 99 <212> TYPE: DNA <213> ORGANISM: Artificial
Sequence <220> FEATURE: <223> OTHER INFORMATION: PCR
product after restriction enzyme digestion <221> NAME/KEY:
misc_feature <222> LOCATION: (1)...(99) <223> OTHER
INFORMATION: n = A,T,C or G <400> SEQUENCE: 705 tcaytagtgc
aaattgnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60
nnnnnnnnnn nnnnncagcc tttatcagga aaattatag 99 <210> SEQ ID NO
706 <211> LENGTH: 97 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION: (1)...(97)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
706 aytagtgcaa attgnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn 60 nnnnnnnnnn nnncagcctt tatcaggaaa attatag 97
<210> SEQ ID NO 707 <211> LENGTH: 13 <212> TYPE:
DNA <213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <400> SEQUENCE: 707 aagtttagat cag 13 <210>
SEQ ID NO 708 <211> LENGTH: 120 <212> TYPE: DNA
<213> ORGANISM: Artificial Sequence <220> FEATURE:
<223> OTHER INFORMATION: PCR product after restriction enzyme
digestion <221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(120) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 708 aattcgtgaa agcagaagtt gtctgataat
cnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnccaaggs 120 <210> SEQ ID
NO 709 <211> LENGTH: 17 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<400> SEQUENCE: 709 aattctgatc taaactt 17 <210> SEQ ID
NO 710 <211> LENGTH: 119 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION:
(1)...(119) <223> OTHER INFORMATION: n = A,T,C or G
<400> SEQUENCE: 710 aagsccttgg nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn nngattatca gacaacttct gctttcacg 119 <210> SEQ ID
NO 711 <211> LENGTH: 117 <212> TYPE: DNA <213>
ORGANISM: Artificial Sequence <220> FEATURE: <223>
OTHER INFORMATION: PCR product after restriction enzyme digestion
<221> NAME/KEY: misc_feature <222> LOCATION: (9)...(90)
<223> OTHER INFORMATION: n = A,T,C or G <400> SEQUENCE:
711 gsccttggnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn
nnnnnnnnnn 60 nnnnnnnnnn nnnnnnnnnn nnnnnnnnnn gattatcaga
caacttctgc tttcacg 117
* * * * *
References