U.S. patent application number 16/947678 was filed with the patent office on 2020-11-26 for identifying variants of interest by imputation.
The applicant listed for this patent is 23andMe, Inc.. Invention is credited to Geoffrey Benton, Arnab Chowdry, Brian Naughton.
Application Number | 20200372974 16/947678 |
Document ID | / |
Family ID | 1000005016281 |
Filed Date | 2020-11-26 |
![](/patent/app/20200372974/US20200372974A1-20201126-D00000.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00001.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00002.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00003.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00004.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00005.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00006.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00007.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00008.png)
![](/patent/app/20200372974/US20200372974A1-20201126-D00009.png)
United States Patent
Application |
20200372974 |
Kind Code |
A1 |
Chowdry; Arnab ; et
al. |
November 26, 2020 |
IDENTIFYING VARIANTS OF INTEREST BY IMPUTATION
Abstract
Processing genetic information comprises: receiving an input
that includes information pertaining to a specific genetic variant;
and identifying, in a database comprising genotype information of a
plurality of candidate individuals, a matching individual imputed
to have the specific genetic variant. The genotype information of
the matching individual corresponding to the specific genetic
variant is not directly assayed.
Inventors: |
Chowdry; Arnab; (Sunnyvale,
CA) ; Benton; Geoffrey; (Cupertino, CA) ;
Naughton; Brian; (Mountain View, CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
23andMe, Inc. |
Sunnyvale |
CA |
US |
|
|
Family ID: |
1000005016281 |
Appl. No.: |
16/947678 |
Filed: |
August 12, 2020 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
15256388 |
Sep 2, 2016 |
10777302 |
|
|
16947678 |
|
|
|
|
13908455 |
Jun 3, 2013 |
|
|
|
15256388 |
|
|
|
|
61689398 |
Jun 4, 2012 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G16B 20/00 20190201;
G16B 40/00 20190201; G16B 50/00 20190201 |
International
Class: |
G16B 40/00 20060101
G16B040/00; G16B 50/00 20060101 G16B050/00; G16B 20/00 20060101
G16B020/00 |
Claims
1. A genetic information processing system, comprising: one or more
computer processors configured to: receive an input that includes
information pertaining to a specific genetic variant; and identify,
in a database comprising genotype information of a plurality of
candidate individuals, a matching individual imputed to have the
specific genetic variant; and one or more memories coupled to the
one or more computer processors and configured to provide the one
or more computer processors with instructions; wherein genotype
information of the matching individual corresponding to the
specific genetic variant is not directly assayed.
2. The system of claim 1, wherein the specific genetic variant is a
variant of unknown significance.
3. The system of claim 1, wherein the one or more computer
processors are further configured to: identify additional matching
individuals imputed to have the specific genetic variant; and
process phenotype information of the matching individuals to
determine an association of the specific genetic variant and a
phenotype.
4. The system of claim 1, wherein the specific genetic variant is
known to be associated with a phenotype, and the one or more
computer processors are further configured to notify the matching
individual of the association.
5. The system of claim 1, wherein to identify the matching
individual includes to perform statistical imputation.
6. The system of claim 1, wherein to identify the matching
individual includes to perform statistical imputation, including
to: establish a statistical model based on genotype information of
a set of reference individuals; and apply a candidate individual's
genotype information to the statistical model to determine whether
the candidate individual is the matching individual who has the
specific genetic variant.
6. The system of claim 6, wherein: the statistical model includes a
haplotype graph; and to apply the candidate individual's genotype
information to the statistical model includes to: identify a likely
genotype sequence based on the candidate individual's genotype and
the haplotype graph; and determine, according to the likely
genotype sequence, whether the candidate individual has the
specific genetic variant.
8. The system of claim 6, wherein: the genotype information of the
set of reference individuals is more densely assayed than genotype
information of the plurality of candidate individuals.
9. The system of claim 1, wherein to identify the matching
individual includes to perform Identity by Descent (IBD)-based
imputation.
10. The system of claim 9, wherein to perform IBD-based imputation
includes to: receive additional genotype information of a proband
who has the specific genetic variant; and determine whether a
candidate individual shares a variant-overlapping IBD region with
the proband.
11. The system of claim 10, wherein to determine whether the
candidate individual shares a variant-overlapping IBD region with
the proband includes to: determine a distance between two adjacent
opposite-homozygous calls in the proband's genotype sequence and
the candidate individual's genotype sequence, one of the two
adjacent opposite-homozygous calls is to the left of the specific
genetic variant and another one of the two adjacent
opposite-homozygous calls is to the right of the specific genetic
variant; and determine whether the distance meets a threshold.
12. The system of claim 1, wherein genetic material of the matching
individual is further assayed to validate whether the matching
individual actually possesses the specific genetic variant.
13. A method of processing genetic information, comprising:
receiving an input that includes information pertaining to a
specific genetic variant; and identifying, using one or more
computer processors and in a database comprising genotype
information of a plurality of candidate individuals, a matching
individual imputed to have the specific genetic variant; wherein
genotype information of the matching individual corresponding to
the specific genetic variant is not directly assayed.
14. The method of claim 13, wherein the specific genetic variant is
a variant of unknown significance.
15. The method of claim 13, further comprising: identifying
additional matching individuals imputed to have the specific
genetic variant; and processing phenotype information of the
matching individuals to determine an association of the specific
genetic variant and a phenotype.
16. The method of claim 13, wherein the specific genetic variant is
known to be associated with a phenotype, and the method further
comprises notifying the matching individual of the association.
17. The method of claim 13, wherein identifying the matching
individual includes performing statistical imputation.
18. The method of claim 13, wherein identifying the matching
individual includes performing statistical imputation, including:
establishing a statistical model based on genotype information of a
set of reference individuals; and applying a candidate individual's
genotype information to the statistical model to determine whether
the candidate individual is the matching individual who has the
specific genetic variant.
19. The method of claim 18, wherein: the statistical model includes
a haplotype graph; and applying the candidate individual's genotype
information to the statistical model includes: identifying a likely
genotype sequence based on the candidate individual's genotype and
the haplotype graph; and determining, according to the likely
genotype sequence, whether the candidate individual has the
specific genetic variant.
20. The method of claim 18, wherein: the genotype information of
the set of reference individuals is more densely assayed than
genotype information of the plurality of candidate individuals.
Description
INCORPORATION BY REFERENCE
[0001] An Application Data Sheet is filed concurrently with this
specification as part of the present application. Each application
that the present application claims benefit of or priority to as
identified in the concurrently filed Application Data Sheet is
incorporated by reference herein in its entirety and for all
purposes.
BACKGROUND OF THE INVENTION
[0002] Genetic researchers often need to study specific genetic
variants to understand their significance. For example, researchers
may be interested in knowing whether a certain genetic variant of
interest (VOI) (e.g., having G/C allele at location 150 on
Chromosome 3) is correlated with a particular phenotype expression
(e.g., having a particular disease). Currently, the interpretation
of specific genetic variants and identification of cohorts with
such variants, particularly variants of unknown significance (VUS)
from whole-genome sequence data, pose substantial challenges in
genetics studies. VUS are so named because their correlations with
specific phenotypes (e.g., certain diseases) are unknown prior to
the studies. VUS are often too rare to be amenable to genome-wide
association studies and thus traditionally have been interpreted
with reference to the primary literature (especially for
high-penetrance or Mendelian mutations) or by computational methods
(e.g., Sorting Intolerant From Tolerant (SIFT), PolyPhen).
[0003] Some large personal genomic information database in
existence can include individuals who actually possess the genetic
variants of interest (VOI). For example, 23andMe, Inc., a personal
genetics service company, has accumulated a large database
comprising data of over 250,000 individuals. The large databases
typically employ genotype data comprising genetic markers to
represent an individual's genome, instead of using full sequence
data. Because the genotype data is usually obtained using chips
that have specific probes assaying selective locations on the
genome, the data is typically not a full sequence and the VOI is
not necessarily directly assayed (for example, an individual's
assayed genotype data may not include specific information about
the person's allele at location 150 on Chromosome 3 because the
chip used for assaying does not have a probe at that location),
making it difficult to study the VOI by directly using information
stored in the large databases.
BRIEF DESCRIPTION OF THE DRAWINGS
[0004] Various embodiments of the invention are disclosed in the
following detailed description and the accompanying drawings.
[0005] FIG. 1 is a functional diagram illustrating a programmed
computer system for performing imputation-based processing of
genetic information in accordance with some embodiments.
[0006] FIG. 2 is a block diagram illustrating an embodiment of a
platform configured to process genetic information based on
imputation.
[0007] FIG. 3 is a flowchart illustrating an embodiment of a
genetic information analysis process.
[0008] FIG. 4 is a flowchart illustrating an embodiment of a
statistical imputation process.
[0009] FIG. 5 is a diagram illustrating an example of a haplotype
graph that is constructed based on a reference collection of
genotype data.
[0010] FIG. 6 is a diagram illustrating an example of VOI
identification based on IBD.
[0011] FIG. 7 is a diagram illustrating an example in which phased
data is compared to identify IBD.
[0012] FIG. 8 is a diagram illustrating an embodiment of another
IBD-based imputation process.
[0013] FIG. 9 is a diagram illustrating example genotype data used
for IBD identification by process 700.
DETAILED DESCRIPTION
[0014] The invention can be implemented in numerous ways, including
as a process; an apparatus; a system; a composition of matter; a
computer program product embodied on a computer readable storage
medium; and/or a computer processor, such as a computer processor
configured to execute instructions stored on and/or provided by a
memory coupled to the computer processor. In this specification,
these implementations, or any other form that the invention may
take, may be referred to as techniques. In general, the order of
the steps of disclosed processes may be altered within the scope of
the invention. Unless stated otherwise, a component such as a
computer processor or a memory described as being configured to
perform a task may be implemented as a general component that is
temporarily configured to perform the task at a given time or a
specific component that is manufactured to perform the task. As
used herein, the term `computer processor` refers to one or more
devices, circuits, and/or processing cores configured to process
data, such as computer program instructions.
[0015] A detailed description of one or more embodiments of the
invention is provided below along with accompanying figures that
illustrate the principles of the invention. The invention is
described in connection with such embodiments, but the invention is
not limited to any embodiment. The scope of the invention is
limited only by the claims and the invention encompasses numerous
alternatives, modifications and equivalents. Numerous specific
details are set forth in the following description in order to
provide a thorough understanding of the invention. These details
are provided for the purpose of example and the invention may be
practiced according to the claims without some or all of these
specific details. For the purpose of clarity, technical material
that is known in the technical fields related to the invention has
not been described in detail so that the invention is not
unnecessarily obscured.
[0016] Imputation-based processing of genetic information is
described. In some embodiments, a database is used to store
genotype information (and optionally other information such as
phenotype information) of a plurality of candidate individuals.
Although the genotype information pertaining to a specific genetic
variant (also referred to as a variant of interest (VOI)) is not
directly assayed (i.e., not directly tested using a genotyping chip
or other genotyping platform) for all the candidate individuals,
matching individuals who are deemed to have the specific genetic
variant are identified using imputation techniques such as
statistical imputation, Identity by Descent (IBD)-based imputation,
or a combination thereof.
[0017] FIG. 1 is a functional diagram illustrating a programmed
computer system for performing imputation-based processing of
genetic information in accordance with some embodiments. As will be
apparent, other computer system architectures and configurations
can be used to perform imputation-based processing of genetic
information. Computer system 100, which includes various subsystems
as described below, includes at least one microprocessor subsystem
(also referred to as a processor or a central processing unit
(CPU)) 102. For example, processor 102 can be implemented by a
single-chip processor or by multiple processors. In some
embodiments, processor 102 is a general purpose digital processor
that controls the operation of the computer system 100. Using
instructions retrieved from memory 110, the processor 102 controls
the reception and manipulation of input data, and the output and
display of data on output devices (e.g., display 118). In some
embodiments, processor 102 includes and/or is used to perform
imputation as described below.
[0018] Processor 102 is coupled bi-directionally with memory 110,
which can include a first primary storage, typically a random
access memory (RAM), and a second primary storage area, typically a
read-only memory (ROM). As is well known in the art, primary
storage can be used as a general storage area and as scratch-pad
memory, and can also be used to store input data and processed
data. Primary storage can also store programming instructions and
data, in the form of data objects and text objects, in addition to
other data and instructions for processes operating on processor
102. Also as is well known in the art, primary storage typically
includes basic operating instructions, program code, data, and
objects used by the processor 102 to perform its functions (e.g.,
programmed instructions). For example, memory 110 can include any
suitable computer-readable storage media, described below,
depending on whether, for example, data access needs to be
bi-directional or uni-directional. For example, processor 102 can
also directly and very rapidly retrieve and store frequently needed
data in a cache memory (not shown).
[0019] A removable mass storage device 112 provides additional data
storage capacity for the computer system 100, and is coupled either
bi-directionally (read/write) or uni-directionally (read only) to
processor 102. For example, storage 112 can also include
computer-readable media such as magnetic tape, flash memory,
PC-CARDS, portable mass storage devices, holographic storage
devices, and other storage devices. A fixed mass storage 120 can
also, for example, provide additional data storage capacity. The
most common example of mass storage 120 is a hard disk drive. Mass
storage 112, 120 generally store additional programming
instructions, data, and the like that typically are not in active
use by the processor 102. It will be appreciated that the
information retained within mass storage 112 and 120 can be
incorporated, if needed, in standard fashion as part of memory 110
(e.g., RAM) as virtual memory.
[0020] In addition to providing processor 102 access to storage
subsystems, bus 114 can also be used to provide access to other
subsystems and devices. As shown, these can include a display
monitor 118, a network interface 116, a keyboard 104, and a
pointing device 106, as well as an auxiliary input/output device
interface, a sound card, speakers, and other subsystems as needed.
For example, the pointing device 106 can be a mouse, stylus, track
ball, or tablet, and is useful for interacting with a graphical
user interface.
[0021] The network interface 116 allows processor 102 to be coupled
to another computer, computer network, or telecommunications
network using a network connection as shown. For example, through
the network interface 116, the processor 102 can receive
information (e.g., data objects or program instructions) from
another network or output information to another network in the
course of performing method/process steps. Information, often
represented as a sequence of instructions to be executed on a
processor, can be received from and outputted to another network.
An interface card or similar device and appropriate software
implemented by (e.g., executed/performed on) processor 102 can be
used to connect the computer system 100 to an external network and
transfer data according to standard protocols. For example, various
process embodiments disclosed herein can be executed on processor
102, or can be performed across a network such as the Internet,
intranet networks, or local area networks, in conjunction with a
remote processor that shares a portion of the processing.
Additional mass storage devices (not shown) can also be connected
to processor 102 through network interface 116.
[0022] An auxiliary I/O device interface (not shown) can be used in
conjunction with computer system 100. The auxiliary I/O device
interface can include general and customized interfaces that allow
the processor 102 to send and, more typically, receive data from
other devices such as microphones, touch-sensitive displays,
transducer card readers, tape readers, voice or handwriting
recognizers, biometrics readers, cameras, portable mass storage
devices, and other computers.
[0023] In addition, various embodiments disclosed herein further
relate to computer storage products with a computer readable medium
that includes program code for performing various
computer-implemented operations. The computer-readable medium is
any data storage device that can store data which can thereafter be
read by a computer system. Examples of computer-readable media
include, but are not limited to, all the media mentioned above:
magnetic media such as hard disks, floppy disks, and magnetic tape;
optical media such as CD-ROM disks; magneto-optical media such as
optical disks; and specially configured hardware devices such as
application-specific integrated circuits (ASICs), programmable
logic devices (PLDs), and ROM and RAM devices. Examples of program
code include both machine code, as produced, for example, by a
compiler, or files containing higher level code (e.g., script) that
can be executed using an interpreter.
[0024] The computer system shown in FIG. 1 is but an example of a
computer system suitable for use with the various embodiments
disclosed herein. Other computer systems suitable for such use can
include additional or fewer subsystems. In addition, bus 114 is
illustrative of any interconnection scheme serving to link the
subsystems. Other computer architectures having different
configurations of subsystems can also be utilized.
[0025] FIG. 2 is a block diagram illustrating an embodiment of a
platform configured to process genetic information based on
imputation. In this example, imputation-based genetic information
processing system 202 (also referred to as the imputation engine)
may be implemented using one or more computers having one or more
processors, one or more special purpose computing appliances, or
any other appropriate hardware, software, or combinations thereof.
The operations of the imputation engine are described in greater
detail below.
[0026] In this example, personal information (including a
combination of genetic information, phenotype information, family
information, and/or population group information) pertaining to a
plurality of candidate individuals is stored in a database 210,
which can be implemented on an integral storage component of the
imputation engine, an attached storage device, a separate storage
device accessible by the imputation engine, or a combination
thereof.
[0027] At least a portion of the database includes genotype data,
specifically genotype data of genetic markers of individuals'
deoxyribonucleic acid (DNA). Examples of such genetic markers
include Single Nucleotide Polymorphisms (SNPs), which are points
along the genome each corresponding to two or more common
variations; Short Tandem Repeats (STRs), which are repeated
patterns of two or more repeated nucleotide sequences adjacent to
each other; and Copy-Number Variants (CNVs), which include longer
sequences of DNA that could be present in varying numbers in
different individuals. Although SNP-based genotype data is
described extensively below for purposes of illustration, the
technique is also applicable to other forms of genotype data such
as STRs, CNVs, etc.
[0028] In this example, genotype data is used to represent the
individuals' genomes instead of full sequence data. In some
embodiments, the genotype data is obtained from DNA samples such as
saliva or blood submitted by individuals. The laboratory analyzes
the samples using a genotyping platform, for example the Illumina
OmniExpress.TM.genotyping chip, which includes probes to assay
allele values for a specific set of SNPs. One genotyping process is
known as hybridization and yields different hybridization intensity
values for each allele. The laboratory assigns genotype values to
the alleles of each SNP by comparing the relative strength of these
intensities. The resulting genotype data is stored in database 210
as a first set of genotype data. Other genotyping techniques can be
used.
[0029] In some embodiments, this first set of genotype data is
referred to as sparsely assayed data. A second set of reference
genotype data (referred to as densely assayed data) is optionally
stored. The reference genotype data includes more densely assayed
genotype data and can be full genome sequence data. Compared with
sparsely assayed data, densely assayed data includes greater
amounts of genotype information as more locations on the
chromosomes are assayed. In some embodiments, densely assayed data
is obtained by combining results from multiple chips each assaying
a different set of markers. For example, one chip assays chromosome
locations 1 and 10, another chip assays chromosome locations 3 and
13, and the results are combined to produce densely assayed data at
chromosome locations 1, 3, 10, and 13. In some embodiments, densely
assayed data is obtained using genotyping platforms that assay a
greater density of genetic markers (e.g., Illumina Omni2.5.TM. or
Omni5.TM.) or sequence the full genome (e.g., ABI SOLiD.TM.) In
some embodiments, a combination of the techniques is used to obtain
the reference data.
[0030] Information pertaining to one or more VOIs, and optionally
additional information such as genotype information of a proband
(i.e., an individual who has the VOI), is input into the imputation
engine. The imputation engine identifies one or more matching
individuals deemed likely to have the VOI even though the genotype
information corresponding to the location of the VOI is not
directly assayed. Examples of imputation technique include building
a statistical model to perform statistical imputation, identifying
Identical by Descent (IBD) regions, or a combination thereof.
[0031] In some embodiments, the imputation engine is a part of a
personal genomic services platform providing a variety of services
such as genetic counseling, ancestry finding, social networking,
etc. In some embodiments, individuals whose data is stored in
database 210 are registered users of a personal genomic service
platform, which provides access to the data and a variety of
personal genetics related services that the individuals have
consented to participate in. Users such as Alice and Bob are
genotyped and their genotype data is stored in database 210. They
access the platform via a network 204 using client devices such as
206 and 208, and interacts with the platform via appropriate user
interfaces (UIs) and applications. A variety of additional actions
are possible. For example, in various embodiments, the correlation
of having the VOI and having a certain phenotype such as a
particular disease is determined; the matching individuals are
recruited for further studies, notified of potential disease risk
due to high correlation of having the genotype and having the
disease, offered tips of possible preventive measures, etc.
[0032] FIG. 3 is a flowchart illustrating an embodiment of a
genetic information analysis process. In some embodiments, process
300 is performed on a system such as 100 or 202.
[0033] At 302, an input that includes information pertaining to a
specific genetic variant (also referred to as the VOI) is received.
For example, the input indicates that the VOI corresponds to G/C
alleles at location 150 on Chromosome 3. Depending on the
application, the significance of the VOI can be unknown or known.
For example, in associative studies, the significance of the VOI is
unknown since the purpose of the study is to determine whether
certain phenotypes are associated (e.g., correlated) with the VOL
Thus, when a cohort of matching individuals is formed, their
phenotype information is further processed to determine whether
there are any associations of phenotypes with the VOL In genetic
counseling or risk analysis applications, the VOI is known to be
associated with certain phenotype expressions and matching
individuals identified as having the VOI are warned of the
potential risk for developing a disease, notified of the known
association, provided with disease prevention tips and other
related information. In some embodiments, the input further
includes additional genotype information of a proband whose genome
includes the VOL The additional genotype information can include a
set of genetic markers or a full genome sequence.
[0034] At 304, imputation is performed to identify, in a database,
one or more matching individuals deemed likely to have the specific
genetic variant. The set of matching individuals is referred to as
a cohort. As discussed above in connection with FIG. 2, the
database comprises genotype information of a plurality of candidate
individuals, and genotype information of the matching individual
pertaining to the specific genetic variant (e.g., genotype
information at the variant location) is not directly assayed. As
will be described in greater detail below, techniques such as
statistical imputation, IBD region identification, and the like are
used to determine which individuals are likely to have the VOI.
[0035] Optionally, at 306, validation is performed to ensure that
the variant is not a private mutation (a mutation present in one
person or family), or a de novo mutation (specifically, a new
mutation in an individual). In some embodiments, once the cohort is
formed, the genetic information of one or more matching individuals
in the cohort is assayed directly to verify whether the mutation is
not private or de novo. In other words, DNAs of one or more
individuals from the cohort are directly assayed (e.g., sequenced
or using chips with a specific probe at the VOI location) to
determine whether they actually possess the VOL If the direct assay
results confirm that the individuals have the VOI, then the VOI is
not a private or de novo mutation; otherwise the VOI is likely a
private or de novo mutation rather than a VOI shared by the
cohort.
[0036] Optionally, at 308, cohort data is output and/or further
processed. For example, statistical analysis such as phenotypic
association of the VOI is performed in some embodiments to
determine whether having the VOI is correlated with having certain
phenotype such as a particular disease or condition.
[0037] Optionally, at 310, matching individuals in the cohort are
notified, preferably via a personal genomic services platform such
as 200. Depending on system configuration, the individuals may be
notified via email, text messages, system messages, or any other
appropriate communication channel. Depending on the application,
the individuals may receive different types of notification. In
embodiments where the cohort is identified for associative studies
of a VOI of unknown significance, the individuals in the cohort are
invited to participate in these studies; in embodiments where users
have previously indicated interest in receiving genetic
counseling/risk analysis information and the VOI is known to be
correlated with certain phenotype, the individuals in the cohort
are notified of the correlation and provided with information such
as the specific risk, prevention tips, etc.
Statistical Imputation
[0038] In some embodiments, imputation includes statistical
imputation. A statistical model such as a haplotype graph is
established based on a set of reference individuals with densely
assayed data. Sparsely assayed genotype data of a candidate
individual (i.e., an individual whose genotype corresponding to the
VOI location is not directly assayed) is applied to the statistical
model to impute whether that individual possesses the VOI.
[0039] The standard SNP-based genotyping technology results in
genotype calls each having two alleles, one from each half of a
chromosome pair. As used herein, a genotype call refers to the
identification of the pair of alleles at a particular locus on the
chromosome. Genotype calls can be phased or unphased. In phased
data, the individual's diploid genotype at a particular locus is
resolved into two haplotypes, one for each chromosome. In unphased
data, the two alleles are unresolved; in other words, it is
uncertain which allele corresponds to which haplotype or
chromosome. Phasing can be done using known phasing tools such as
BEAGLE. Either phased data or unphased data can be used in the
imputation process; for purposes of illustration examples using
phased data are discussed with respect to the statistical
imputation process.
[0040] Examples of sparsely assayed data of candidate individuals
and densely assayed data of reference individuals are illustrated
in Tables 1 and 2, respectively.
TABLE-US-00001 TABLE 1 Genotype Data Phenotype Data ID 100 200 300
400 . . . Disease X? Disease Y? . . . U1 G T A C . . . Yes No U2 A
C G G No No U3 G A C A No Unknown . . . . . . . . . . . . . . . . .
. . . .
[0041] Table 1 illustrates an example data set of sparsely assayed
genotype data and phenotype data of candidate individuals. In some
embodiments, the data is obtained using assaying chips or other
genotyping hardware, and is stored in a database such as 210. In
this example, data for an individual is stored in a row. The
genotype data of each individual comprises a set of SNPs with
values such as A, C, G, or T at specific locations on a chromosome.
For purposes of simplicity, this example assumes that phased data
is used, and unphased data can be processed using known phasing
techniques such as BEAGLE to obtain phased data. For purposes of
illustration, the genotype data includes N SNPs, shown to be
assayed at every 100.sup.th location on the chromosome (i.e., at
locations 100, 200, 300, and 400); other locations may be used in
other implementations. This set of genotype data is sparsely
assayed data. Phenotype data such as whether the individual has a
particular disease or condition is also stored. Other data formats
may be used in other embodiments.
[0042] To perform statistical imputation, a reference data set of
densely assayed data is used to construct a statistical model
(e.g., a haplotype graph) used to determine likely genotype
sequences for the candidate individuals. Table 2 illustrates an
example data set of reference data. For purposes of illustration,
the reference data includes L SNPs, assayed at every 20.sup.th
location at locations 20, 40, 60, 80, 100, etc. Other locations may
be used in other embodiments. In some embodiments, full genome
sequences are used. The number of reference individuals in the
densely assayed reference data set is typically fewer than the
number of candidate individuals in the sparsely assayed data set.
For example, there can be 100,000 or more individuals in the
sparsely assayed data set, but only 1000 in the densely assayed
data set.
TABLE-US-00002 TABLE 2 Genotype Data ID 20 40 60 80 100 120 140 160
. . . R1 A A G A C T T G . . . R2 A G T C C T A G . . . R3 G T A A
T C G C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . .
[0043] FIG. 4 is a flowchart illustrating an embodiment of a
statistical imputation process. In some embodiments, process 400 is
used to implement step 304 of process 300.
[0044] At 402, a statistical model is established based on the
reference data comprising densely assayed genotype data. In some
embodiments, the statistical model includes a haplotype graph that
represents the genotype data of the reference population. The
reference data includes genotype information corresponding to the
VOI location.
[0045] FIG. 5 is a diagram illustrating an example of a haplotype
graph that is constructed based on a reference collection of
genotype data. In this example, the reference collection of
genotype data includes a set of L (L=6 in this case) genetic
markers (e.g., SNPs) measured at locations 20, 40, 60, 80, 100, and
120 on a particular chromosome. In other embodiments, different L
and different locations can be used. The reference data includes
directly assayed genotype information corresponding to the VOI
location. For example, if the VOI is located at location 80 of the
chromosome, the genotype of the reference individuals' genome at
this location is directly assayed and the information is included
in the reference genotype data.
[0046] As shown, haplotype graph 500 is a Directed Acyclic Graph
(DAG) having nodes (e.g., 504) and edges (e.g., 506). To construct
a haplotype graph, each reference individual's genotype data forms
a path. Identical portions of the paths are combined to form nodes
and edges of the graph. The haplotype graph starts with a single
node (the "begin state") and ends on a single node (the "accepting
state"), and the intermediate nodes correspond to the states of the
markers at respective gene loci. There are a total of L+1 levels of
nodes from left to right. An edge, e, represents the set of
haplotypes whose path from the initial node to the terminating node
of the graph traverses e. The possible paths define the haplotype
space of possible genotype sequences. For example, in haplotype
graph 500, a path 502 corresponds to the genotype sequence
"GTTCAC." In the example shown, there are four possible
paths/genotype sequences in the haplotype space shown in this
diagram ("ACGCGC," "ACTTAC," "GTTCAC," and "GTTTGG"). Given a
greater number of reference individuals and longer genotype
sequences, the resulting haplotype graph would be a more complex
structure having a greater number of nodes and possible paths.
[0047] Each edge in the haplotype graph is associated with a
probability computed based on the reference collection of densely
assayed genotype data. In this example, the reference collection is
comprised of genotype data from 1000 individuals, of which 400 have
the "A" allele at the first position (in this case, location 20 on
the chromosome), and 600 have the "G" allele at the first position.
Accordingly, the probability associated with edge 508 is 400/1000
and the probability associated with edge 510 is 600/1000. All of
the first 400 individuals have the "C" allele at the second
position (location 40 on the chromosome), giving edge 512 a
probability of 400/400. All of the next 600 individuals who had the
"G" allele at the first position have the "T" allele at the second
position, giving edge 514 a probability of 600/600, and so on. The
probabilities associated with the respective edges are labeled in
the diagram. The probability associated with a specific path is
expressed as the product of the probabilities associated with the
edges included in the path. For example, the probability associated
with path 502 ("GTTCAC") is computed as:
P(h)=(600/1000)(600/600)(600/600)(50/600)(350/350)(450/450)=0.05
[0048] Returning to FIG. 4, at 404, genotype data of a candidate
individual whose genotype at the position corresponding to the VOI
is not directly assayed is input into the statistical model. In
some embodiments, sparsely genotyped data of the candidate
individual is applied as inputs to the haplotype graph.
[0049] At 406, a likely genotype sequence is identified based on
the candidate individual's genotype data and the statistical model.
In some embodiments, at least a portion of the sparsely genotyped
data (e.g., a portion that overlaps the VOI location) is compared
with paths on the haplotype graph to find a most likely path (i.e.,
a likely genotype sequence). Referring again to FIG. 5 for an
example: suppose that the only assayed SNP in the individual's
genome corresponds to the SNP at the fifth position of the
haplotype graph (location 100 on the chromosome) and is "A."
Accordingly, there are two possible paths identified for this
individual: "GTTCAC" (path 502) and "ACTTAC." The probability
associated with the path "ACTTAC" is computed as:
P(h)=(400/1000)(400/400)(300/400)(300/300)(350/350)(450/450)=0.03
[0050] Further, as discussed above, the probability associated with
path "GTTCAC" is 0.05. Thus, it is more likely that the
individual's genotype sequence follows the "ACTTAC" path instead of
the "GTTCAC" path. In other words, the individual is imputed to
have A, C, T, T, and C at locations 20, 40, 60, 80, and 120 of the
chromosome even though only the SNP at location 100 is directly
assayed.
[0051] In some embodiments, to determine the most likely path, all
possible paths are identified based on the assayed SNPs and the
probability associated with each possible path is determined to
find the most likely one. In some embodiments, instead of computing
all possible paths, a dynamic programming technique is used to
improve computational efficiency. In some embodiments, instead of
the most likely full path, a most likely portion of a path (e.g., a
portion of a genotype sequence that overlaps the VOI location) is
determined. In some embodiments, the portion size is empirically
determined and depends on the length of the haplotypes, which can
vary depending on the amount of recombination in a region and the
relative age of the variant.
[0052] At 408, the likely genotype sequence is optionally stored to
facilitate VOI lookups. Future VOI determinations will not require
the sequence to be recomputed; instead, the stored likely genotype
sequence is looked up, and performance is improved.
[0053] At 410, whether the candidate individual has the VOI is
determined according to the likely genotype sequence. Referring
again to FIG. 5, suppose that the VOI is "C" at location 80 of the
chromosome (the fourth position on the haplotype graph). In the
event that the most likely path of a candidate individual is
"ACTTAC," the variant at location 80 (the fourth position of the
sequence) is imputed to be "T" and the individual is imputed not to
have the VOL In the event that the most likely path of the
individual follows "GTTCAC" (possibly determined as a result of
other matches to locations not shown on the haplotype graph of FIG.
5, such as locations 200, 300, etc.), the variant at location 80 is
imputed to be "C" and the individual is imputed to have the
VOI.
[0054] In some embodiments, 402-410 are repeated (e.g., executed
multiple times sequentially or in parallel) for a plurality of
individuals whose genome is sparsely assayed. Individuals deemed to
likely have the VOI form a cohort.
[0055] As discussed above, the reference individuals' genotype
information is more densely assayed than that of the candidate
individuals. In some embodiments, the haplotype graph is used to
help determine locations in the candidate individuals' genome to
assay. For example, referring to FIG. 5, locations where the
haplotype graph branches out (e.g., location 20 with the A/G
branch, location 60 with the G/T branch, location 80 with the C/T
branch) have more information and would improve the accuracy of
identification of the likely paths of the candidate individuals.
Thus, if the candidate individuals have not been assayed already or
if re-assaying is permitted, the system can opt to assay the
candidate individuals at locations 20, 60, and/or 80 instead of
100.
Identity by Descent (IBD)-based imputation
[0056] In some embodiments, imputation includes identifying IBD
regions between a proband and a candidate individual. IBD-based
imputation does not require a reference set of densely assayed
genotype data.
[0057] Because of recombination and independent assortment of
chromosomes, the autosomal DNA and X chromosome DNA (collectively
referred to as recombining DNA) from the parents are shuffled at
the next generation, with small amounts of mutation. Relatives
(i.e., people who descended from the same ancestor) will share long
stretches of genome regions where their recombining DNA is
completely or nearly identical. Such regions are referred to as
"Identity (or Identical) by Descent" (IBD) regions because they
arose from the same DNA sequences in an earlier generation. In some
embodiments, individuals in the database that share a
variant-overlapping IBD region with the proband are identified. A
variant-overlapping IBD region is an IBD region that overlaps the
location where the VOI is found.
[0058] In some embodiments, the determination of IBD regions
includes comparing the DNA markers (e.g., SNPs, STRs, CNVs, etc.)
of two individuals. The standard SNP based genotyping technology
results in genotype calls each having two alleles, one from each
half of a chromosome pair. As used herein, a genotype call refers
to the identification of the pair of alleles at a particular locus
on the chromosome.
[0059] FIG. 6 is a diagram illustrating an example of VOI
identification based on IBD. In this example, the phase haplotype
of the proband, X, is represented by line 602. A particular VOI in
X's genome is identified at location 604. The sequence of a
reference genotype overlapping this location is "AGTCCG," and the
sequence of X's genome is "AGTACG" (the second A being the VOI).
X's genome can be fully sequenced, or genotyped and in addition
have the variant at 604 specifically identified. Note that the
technique of identifying the variant at 604 need not be the same as
assaying the SNP variant used to compute IBD between X and other
individuals in the database.
[0060] In this example, Alice, Bob, Charlie, and Dora are candidate
individuals whose genotype information is stored in the database.
The number of candidate individuals in the database can be much
greater in practice. Note that the genotype information of these
candidate individuals at location 604 is not directly assayed. The
chips used to assay their DNA samples in this case produce SNPs at
other locations of the individuals' genome. The SNPs are
represented as dots on the line.
[0061] It is determined whether individuals in the database (such
as Alice, Bob, Charlie, and Dora) have DNA segments overlapping the
variant location 604 that is IBD with respect to X's genome. IBD
identification can be performed using existing IBD identification
techniques such as fastIBD. In this case, although Alice shares an
IBD region 606 (shown as the shaded box) with X, region 606 does
not overlap the variant location; Bob shares IBD regions 608 and
610 with X, but neither region overlaps the variant location;
Charlie shares IBD region 612 with X, and this region overlaps the
variant location; Dora shares IBD region 614 with X, and this
region also overlaps the variant location. These IBD segments are
typically short (e.g., <10 centimorgans (cM)) and often belong
to distant relatives who share only one segment with the proband.
Since the IBD regions are supposed to be identical between
individuals, although the sequence information for location 604 is
not necessarily known for individuals Charlie and Dora, because
they share IBD regions with X that overlap the variant location, it
is imputed that Charlie and Dora would also have the same variant
at location 604 (in this case, "AGTACG"). In this case, Charlie and
Dora are deemed to be matching individuals since they both share
variant-overlapping IBD regions with X (and therefore are assumed
to have the same VOI).
[0062] In some embodiments, the genotype and/or phenotype
information of Charlie and Dora is output and can be used for a
variety of purposes. For example, Charlie and Dora can be included
in a cohort for statistical analysis, testing for phenotypic
association with this particular variant, etc. For instance, the
technique can be used to test rare mutations (particularly ones
that are not present on the genotyping chip and therefore not
genotyped) in cancer genes to determine (or validate existing
theories on) whether specific mutations are associated with
increased odds of getting a particular type of cancer or are benign
polymorphisms that are not associated with increased risk of
getting that type of cancer.
[0063] In some embodiments, further validation is performed to
confirm that the variant is not a private mutation or a de novo
mutation. Using the example above, once it is established that
Charlie and Dora are in the cohort, if further assaying of Charlie
and Dora's genetic materials shows that neither has the same
variant as X at location 604, then it is likely that X's variant is
private or de novo. If, however, Charlie and/or Dora have the same
variant, then it is unlikely that X's variant is private or de
novo.
[0064] In some embodiments, sequences of SNPs are stored in
dictionaries using a hash-table data structure for the ease of
comparison. FIG. 7 is a diagram illustrating an example in which
phased data is compared to identify IBD. The sequences are split
along pre-defined intervals into non-overlapping words. Other
embodiments may use overlapping words. Although a preset length of
3 is used for purposes of illustration in the example shown, many
implementations may use words of longer lengths (e.g., 100). Also,
the length does not have to be the same for every location. In FIG.
7, on Alice's chromosome pair 1, phased haplotype 902 is
represented by words AGT, CTG, CAA, . . . and phased haplotype 904
is represented by CGA, CAG, TCA, . At each location, the words are
stored in a hash table that includes information about a plurality
of individuals to enable constant retrieval of which individuals
carry matching haplotypes. Similar hash tables are constructed for
other sequences starting at other locations. To determine whether
Bob's chromosome pair 1 shares any IBD with Alice's, Bob's
sequences are processed into words at the same locations as
Alice's. Thus, Bob's haplotype 906 yields CAT, GAC, CCG, . . . and
haplotype 908 yields AAT, CTG, CAA, . . . Every word from Bob's
chromosomes is then looked up in the corresponding hash table to
check whether any other users have the same word at that location
in their genomes. In the example shown, the second and third words
of haplotype 908 match second and third words of Alice's haplotype
902. This indicates that SNP sequence CTGCAA is present in both
chromosomes and suggests the possibility of IBD sharing. If enough
matching words are present in close proximity to each other, the
region would be deemed IBD.
[0065] FIG. 8 is a diagram illustrating an embodiment of another
IBD-based imputation process. Process 700 may be used to implement
304 of process 300 and is applicable to unphased data.
[0066] The genotype call at a particular SNP location may be a
heterozygous call with two different alleles or a homozygous call
with two identical alleles. A heterozygous call is represented
using two different letters such as AB that correspond to different
alleles. Some SNPs are biallelic SNPs with only two possible states
for SNPs. Some SNPs have more states, e.g., triallelic. Other
representations are possible.
[0067] In this example, the letter "A" is selected to represent an
allele with base A and the letter "B" represents an allele with
base G at the SNP location. Other representations are possible. A
homozygous call is represented using a pair of identical letters
such as AA or BB. The two alleles in a homozygous call are
interchangeable because the same allele came from each parent. When
two individuals have opposite-homozygous calls at a given SNP
location, or, in other words, one person has alleles AA and the
other person has alleles BB, it is very likely that the region in
which the SNP resides does not have IBD since different alleles
came from different ancestors. If, however, the two individuals
have compatible calls, that is, both have the same homozygotes
(i.e., both people have AA alleles or both have BB alleles), both
have heterozygotes (i.e., both people have AB alleles), or one has
a heterozygote and the other a homozygote (i.e., one has AB and the
other has AA or BB), there is some chance that at least one allele
is passed down from the same ancestor and therefore the region in
which the SNP resides is IBD. Further, based on statistical
computations, if a region has a very low rate of
opposite-homozygote occurrence over a substantial distance, it is
likely that the individuals inherited the DNA sequence in the
region from the same ancestor and the region is therefore deemed to
be an IBD region.
[0068] At 702, consecutive opposite-homozygous calls in two
individuals' SNPs (e.g., the SNPs of the proband and the SNPs of a
candidate individual) are identified. The consecutive
opposite-homozygous calls can be identified by serially comparing
individual SNPs in the individuals' SNP sequences or in parallel
using bitwise operations. At 704, the distance between two adjacent
opposite-homozygous calls located on either side of the VOI (e.g.,
one to the left side of the VOI and one to the right side of the
VOI) is determined. The distance may be physical distance measured
in the number of base pairs or genetic distance accounting for the
rate of recombination. At 706, it is determined whether the
distance between the opposite-homozygous calls exceeds a threshold.
In some embodiments, the threshold value is set to 10 cM. If the
threshold is exceeded, the region between the calls is determined
to be an IBD region and VOI is imputed to exist in the candidate
individual's genome.
[0069] In some embodiments, a tolerance for genotyping error can be
built by allowing some low rate of opposite homozygotes when
calculating an IBD segment. In some embodiments, the total number
of matching genotype calls is also taken into account when deciding
whether the region is IBD. For example, a region is optionally
further examined where the distance between consecutive opposite
homozygous calls is just below the 10 cM threshold. If a large
enough number of genotype calls within that interval match exactly,
the interval is still deemed IBD and VOI is deemed to exist.
[0070] FIG. 9 is a diagram illustrating example genotype data used
for IBD identification by process 700. 802 and 804 correspond to
the SNP sequences of X and Bob, respectively. In this example, X is
the proband and Bob is a candidate individual. At location 806, the
alleles of X and Bob are opposite-homozygotes, suggesting that the
SNP at this location resides in a non-IBD region. Similarly, at
location 808, the opposite-homozygotes suggest a non-IBD region. At
location 810, however, both pairs of alleles are heterozygotes,
suggesting that there is potential for IBD. Similarly, there is
potential for IBD at location 812, where both pairs of alleles are
identical homozygotes, and at location 814, where X's pair of
alleles is heterozygous and Bob's is homozygous. If there is no
other opposite-homozygote between 806 and 808 and there are a large
number of compatible calls between the two locations (e.g., the
number of compatible calls meeting a predefined threshold), it is
then likely that the region between 806 and 808 is an IBD region.
Since the VOI is located between the adjacent opposite-homozygous
pair 806 and 808, Bob is imputed to have the VOI due to sharing an
IBD region with X.
[0071] In various embodiments, the effects of genotyping error are
accounted for and corrected. In some embodiments, certain genotyped
SNPs are removed from consideration if there are a large number of
Mendelian errors when comparing data from known parent/offspring
trios. In some embodiments, SNPs that have a high no-call rate or
otherwise failed quality control measures during the assay process
are removed. In some embodiments, in an IBD segment, an occasional
opposite-homozygote is allowed if there is sufficient
opposite-homozygotes-free distance (e.g., at least 3 cM and 300
SNPs) surrounding the opposite-homozygote.
[0072] In some embodiments, multiple imputation processes are
pipelined to improve performance. For example, a statistical
imputation process is performed first. If the process fails to find
a match because, for example, the VOI does not exist in the
haplotype graph, then an IBD-based imputation process is performed.
Alternatively, IBD-based imputation can be performed first, and if
no VOI is found, statistical imputation is performed next. In some
embodiments, both statistical imputation and IBD-based imputation
are performed regardless of success or failure of an individual
process. Results from the processes (e.g., matching individuals A,
B, and C found using a statistical imputation process and matching
individuals A, D, and E found using an IBD-based process) are
combined to form the cohort (e.g., A, B, C, D, and E).
[0073] Identifying in a database matching individuals who are
imputed to have a specific genetic variant has been described.
Imputation allows interested parties (e.g., researchers, service
providers) to leverage genetic information in large databases to
find matching individuals, even though the database does not
necessarily have directly assayed information of the specific
variant.
[0074] Although the foregoing embodiments have been described in
some detail for purposes of clarity of understanding, the invention
is not limited to the details provided. There are many alternative
ways of implementing the invention. The disclosed embodiments are
illustrative and not restrictive.
* * * * *