U.S. patent application number 14/543160 was filed with the patent office on 2015-05-21 for lab on chip image analysis.
The applicant listed for this patent is STMicroelectronics, S.r.l.. Invention is credited to Maria Eloisa CASTAGNA, Sabrina CONOCI, Dario RUSSO, Massimo Orazio SPATA.
Application Number | 20150140562 14/543160 |
Document ID | / |
Family ID | 49920533 |
Filed Date | 2015-05-21 |
United States Patent
Application |
20150140562 |
Kind Code |
A1 |
CONOCI; Sabrina ; et
al. |
May 21, 2015 |
LAB ON CHIP IMAGE ANALYSIS
Abstract
An optical analyzer, configured to receiving a cartridge for
biochemical analyses including a plurality of wells, and carrying
out fluorescence analysis of the cartridge, in particular of the
wells. The analyzer comprises a control unit (microprocessor and
memory) configured to: (a) acquiring an image associated to the
fluorescence emitted by the cartridge; (b) identifying, in the
image, pixels belonging to boundary lines of the wells by
generating a black and white binary image; (c) recognizing, between
the geometrical shapes identified in step (b), the ones that best
approximate a reference geometrical shape identifying an ideal
well; (d) acquiring, from the starting image, values of light
intensity only in portions of the image itself that in turn
correspond to the regions recognized in step (c); and (e)
evaluating a state of advance and/or a result of the biochemical
analysis on the basis of the values of light intensity acquired in
step (d).
Inventors: |
CONOCI; Sabrina;
(Tremestieri Etneo, IT) ; CASTAGNA; Maria Eloisa;
(Catania, IT) ; SPATA; Massimo Orazio; (Catania,
IT) ; RUSSO; Dario; (Cerano, IT) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
STMicroelectronics, S.r.l. |
Agrate Brianza |
|
IT |
|
|
Family ID: |
49920533 |
Appl. No.: |
14/543160 |
Filed: |
November 17, 2014 |
Current U.S.
Class: |
435/6.11 ;
382/129; 435/6.12 |
Current CPC
Class: |
B01L 2200/12 20130101;
Y10T 29/49885 20150115; C12Q 1/6851 20130101; B01L 2300/163
20130101; B01L 2300/0819 20130101; B01L 2200/10 20130101; G06T
7/0012 20130101; G06T 2207/10064 20130101; C12Q 1/6825 20130101;
G01N 21/253 20130101; B01L 2300/0816 20130101; G06T 2207/30072
20130101; G01N 21/01 20130101; G01N 21/0303 20130101; B01L
2300/0851 20130101; B01L 2300/168 20130101; B01L 2300/161 20130101;
B01L 2300/0829 20130101; C12Q 1/686 20130101; G06T 7/0014 20130101;
B01L 2300/1827 20130101; G06T 2207/20061 20130101; G01N 21/51
20130101; G06T 2207/20024 20130101; B01L 3/50851 20130101; G06K
9/0014 20130101; G01N 21/6452 20130101; B01L 7/52 20130101; B01L
2300/0887 20130101; B01L 2300/0893 20130101 |
Class at
Publication: |
435/6.11 ;
382/129; 435/6.12 |
International
Class: |
C12Q 1/68 20060101
C12Q001/68; G06T 7/00 20060101 G06T007/00; G06K 9/00 20060101
G06K009/00 |
Foreign Application Data
Date |
Code |
Application Number |
Nov 20, 2013 |
IT |
TO2013A000940 |
Claims
1. A method of acquiring images from an optical process in a
plurality wells of a microchip and processing such images in a way
to obtain a quantitative result, said method comprising: a)
acquiring from an image sensor a first image representing a first
emitted light radiation from a plurality of wells in a microchip;
b) identifying in said first image pixels belonging to a boundary
line of each of said plurality of wells; c) generating a binary
image wherein a first value of light intensity is attributed to
pixels belonging to said boundary lines, and a second value of
light intensity, different from said first value of light
intensity, is attributed to the remaining pixels of said binary
image, whereby said pixels having the first value of light
intensity form, in said binary image, one-dimensional or
two-dimensional geometrical shapes; d) recognizing among said
geometrical shapes, those shapes that best approximate a reference
geometrical shape identifying an ideal well; e) generating a second
image wherein said reference geometrical shape is associated with
each geometrical shape defined by the pixels having the first value
of brightness only in the case where said recognizing step has
yielded a positive result; f) acquiring from said first image,
values of light intensity only at portions of the first image
corresponding to respective portions of the second image internally
delimited the reference geometrical shapes; g) determining a result
of said biochemical or chemical process in each well on the basis
of said values of light intensity detected in each ideal well; h)
repeating steps a-g at successive instants in time; and, i)
calculating a quantitative biochemical result based on the results
of steps (g) and (h).
2. The method according to claim 1, wherein identifying step (b) is
performed using a Canny edge-detection filter.
3. The method according to claim 1, wherein recognizing step (d) is
performed using a Hough or Radon filter.
4. The method according to claim 1, wherein determining step (g)
comprises calculating an arithmetic average of the values of light
intensity detected, and supplying a result of the evaluation based
on the mean value of intensity calculated.
5. The method according to claim 1, wherein said optical process is
a PCR analysis.
6. The method according to claim 5, wherein calculating step i)
comprises: a) tracing a curve of interpolation of the values of
light intensity acquired for each first image at the successive
instants in time; and, b) concluding that an amplification of said
PCR analysis is successful if the curve is a sigmoid, otherwise
concluding an indication that said amplification of said PCR
analysis failed.
7. The method according to claim 1, wherein calculating step i)
comprises: a) calculating a value of light intensity emitted by
each well by performing a sum of the values of light intensity of a
number of pixels of each well and a division by said number of
pixels; and, b) comparing the calculated value of light intensity
with a threshold, thus determining a value of fluorescence each of
said wells.
8. The method according to claim 5, further comprising carry out a
plurality of PCR thermal cycles, wherein steps (a)-(g) are carried
out during each PCR thermal cycle.
9. The method according to claim 5, further comprising the steps
of: a) supplying a biological specimen to be analyzed; b)
extracting DNA from said biological specimen; c) filling said wells
with probes, reagents for PCR, primers and said DNA; d) carrying
out a plurality of thermal cycles each comprising a step of heating
to denature said DNA; a step of annealing said primers to said DNA,
a step of extending said primers to make amplified DNA and a step
of hybridization of said amplified DNA to said probes.
10. The method of claim 5, wherein identifying step b is performed
using a Canny edge-detection filter.
11. The kit according to claim 5, wherein recognizing step d is
performed using a Hough or Radon filter.
12. The method of claim 5, wherein acquiring step f comprises
calculating the arithmetic average of the values of light intensity
detected, and supplying a result of the evaluation based on the
mean value of intensity calculated.
13. The method according to claim 5, wherein the identifying step b
is performed using a Canny edge-detection filter; the recognizing
step d is performed using a Hough or Radon filter; and acquiring
step f comprises calculating the arithmetic average of the values
of light intensity detected, and supplying a result of the
evaluation based on the mean value of intensity calculated.
14. A method for carrying out a biochemical process, comprising the
following steps in order: a) activating a light source to
illuminate a plurality of wells with a excitation light radiation
having a first wavelength (.lamda..sub.E1) such that an emitted
light radiation (.lamda..sub.E2), having a second wavelength, is
emitted by said wells in response to said excitation light
radiation; b) acquiring a first image representing said emitted
light radiation from said plurality of wells; c) identifying pixels
belonging to boundary lines of each of said wells, generating a
binary image wherein a first value of light intensity is attributed
to pixels belonging to said boundary lines, and a second value of
light intensity, different from said first value of light
intensity, is attributed to the remaining pixels of said binary
image, said pixels having the first value of light intensity
defining one-dimensional or two-dimensional geometrical shapes in
said binary image; d) recognizing, among said geometrical shapes,
the ones that best approximate a reference geometrical shape
identifying an ideal well; e) generating a second image in which
said reference geometrical shape is associated to each geometrical
shape defined by the pixels having the first value of brightness
only in the case where the recognition has yielded a positive
result; f) acquiring, from the first image, values of light
intensity only at portions of the first image corresponding to
respective portions of the second image inside respective reference
geometrical shapes associated to respective geometrical shapes
formed by the pixels having the first value of brightness; and, g)
determining a result of said biochemical analysis on the basis of
said values of light intensity detected; and h) repeating steps a-g
at successive instants in time so as to generate a quantitative
biochemical result.
15. The method according to claim 14, further comprising the
preceding steps of: a) supplying a biological specimen to be
analyzed; b) extracting DNA from said biological specimen; c)
filling said wells with probes, PCR reagents, a pair of primers,
and said extracted DNA; and, d) carrying out a plurality of thermal
cycles comprising a denaturing step, an annealing step, a primer
extension step, and a hybridization step which generates said
emitted light radiation.
16. The method according to claim 15, wherein said successive
instants in time occur during said hybridization step.
17. The method of claim 15, wherein identifying step b is performed
using a Canny edge-detection filter.
18. The method of claim 15, wherein recognizing step d is performed
using a Hough or Radon filter.
19. The method of claim 14, wherein acquiring step f comprises
calculating the arithmetic average of the values of light intensity
detected, and supplying a result of the evaluation based on the
mean value of intensity calculated.
20. The method of claim 14, wherein the identifying step b is
performed using a Canny edge-detection filter; the recognizing step
d is performed using a Hough or Radon filter, and acquiring step f
comprises calculating the arithmetic average of the values of light
intensity detected, and supplying a result of the evaluation based
on the mean value of intensity calculated.
Description
[0001] This application claims priority to TO2013A000940, filed
Nov. 20, 2013, and incorporated by reference herein in its entirety
for all purposes.
FIELD OF THE DISCLOSURE
[0002] The present disclosure relates to methods, systems or kits
for the quantitative analysis of reactions occurring in a chip or
other microfluidic device, in particular to a method of image
analysis.
BACKGROUND OF THE DISCLOSURE
[0003] As is known, the analysis of nucleic acids requires
preliminary steps of preparation of a specimen of biological
material, amplification of the nucleic material contained therein,
and hybridization of individual target strands or reference
strands, corresponding to the sequences sought. At the end of these
preparatory steps, the specimen may also be examined to check
whether the amplification has been carried out correctly.
[0004] According to the methodology referred to as "real-time
polymerase-chain-reaction" or "PCR", the DNA is amplified through
thermal cycles appropriately selected, and the progress of the
amplification is detected and monitored by fluorescence during the
entire process.
[0005] Various methods and apparatuses for inspection of an optical
type are known for this purpose. In particular, the methods and
apparatuses of an optical type are frequently based upon the
phenomenon of fluorescence. The amplification reactions are
conducted so that the nucleic acid strands, contained in a
recognition chamber provided in a support, include fluorescent
molecules or fluorophores. The support is exposed to a light source
having an appropriate spectrum of emission, such as to excite the
fluorophores. In turn, the excited fluorophores emit a secondary
radiation at a wavelength of emission higher than the peak of the
excitation spectrum. The light emitted by the fluorophores is
collected and detected by an optical sensor. In order to eliminate
the background light radiation, which represents a source of noise,
the optical sensor is provided with bandpass filters centered at
the wavelength of emission of the fluorophores.
[0006] The detection of different substances in the same specimen
normally requires the use of distinct fluorophores, which have
respective excitation and emission wavelengths. Light sources with
different emission spectra are then used in succession, for
analyzing the responses in the excitation and emission bands of
each fluorophore.
[0007] PCR analyzers designed for being used for optical reading of
the specimens are described in the documents Nos. US20120170608 and
US20130004954. These analyzers are designed to read supports
provided with a relatively small number of wells (in particular,
six wells) containing the specimens for being analyzed, and each
well has relatively large dimensions; i.e., it has a square or
round shape with a side or diameter between 3 and 4 mm.
[0008] Other biochemical and chemical analyses may be similar,
sometimes substituting an antibody for the detection of proteins,
or other ligands for the detection of other chemicals. However,
many are also amendable to a fluorescence-based analysis.
[0009] Known systems present some limitations. In particular, the
composition of the surface of the chip typically used for analysis
generates undesirable reflections of the optical source used for
illuminating the support during fluorescence analysis, generating
background noise that raises the detection-sensitivity
threshold.
[0010] Furthermore, the present applicant has found that the
systems described in the documents Nos. US20120170608 and
US20130004954 do not enable optical reading of supports of silicon
having wells of dimensions smaller than those envisaged by the
known art described therein (for example, square wells with a side
equal to or smaller than 1 mm). Images, typically low-resolution
ones (e.g., 300.times.200 pixels), made by such systems create
difficulty in the discrimination of the individual wells and in
selective acquisition of the brightness of the fluorescence emitted
by the wells. In conclusion, a reading based upon analysis of the
fluorescence of said supports is, in effect, impracticable for
values of fluorescence below a certain threshold that are typical
of a real-time PCR, antibody detection or otherwise.
[0011] Analysis on such small volumes envisages use of
high-resolution image-acquisition systems, or a modification of the
light-emission sources and an increase of the cost and size of the
reading apparatus, which is undesirable as increasing cost.
[0012] Thus, what is needed in the art are better devices and
methods for the collection and analysis of image date when
available from a tiny well or other site.
SUMMARY OF THE DISCLOSURE
[0013] The aim of the present disclosure is to provide a kit and
devices for biochemical analyses and a method for carrying out a
biochemical process that will enable the limitations described
above for being overcome and, in particular, that will enable
reduction of the risk of reading errors during said analysis.
[0014] According to the present disclosure, a kit and device for
biochemical analyses and a method for carrying out a biochemical
process are provided, as defined in the annexed claims.
[0015] For example, an optical analyzer is configured to carry out
fluorescence analysis of a plurality of small wells. The analyzer
comprises a control unit (microprocessor and memory) configured to:
(a) acquire an image associated with the fluorescence emitted by
the cartridge; (b) identify, in the image, pixels belonging to
boundary lines of the wells by generating a black and white binary
image; (c) recognize, between the geometrical shapes identified in
step (b), the ones that best approximate a reference geometrical
shape identifying an ideal well; (d) acquire, from the starting
image, values of light intensity only in portions of the image
itself that in turn correspond to the regions recognized in step
(c); and, (e) evaluate a state of advance and/or a result of the
biochemical analysis on the basis of the values of light intensity
acquired in step (d).
[0016] Advantageously, the interwell regions are coated with an
anti-reflective layer, such as described herein, providing less
interference from random reflections off the cartridge. For
example, the effective thickness h of the passivation layer outside
the wells is thus chosen between .lamda..sub.E cos .theta./(4n)-10%
and .lamda..sub.E cos .theta./(4n)+10%, preferably .lamda..sub.E
cos .theta./(4n), whereas inside the wells is chosen different from
.lamda..sub.E/(4n), for example less than .lamda.E/(4n), where n is
the index of refraction of the passivation layer and .theta. is the
angle formed with the normal to the plane of incidence of the
radiation and .lamda..sub.E is the wavelength of exciting
light.
[0017] The disclosure includes the following embodiments, in any
combination of one or more thereof: [0018] A method of acquiring
images from an optical process in a plurality wells of a microchip
and processing such images in a way to obtain a quantitative
result, said method comprising: [0019] a) acquiring from an image
sensor a first image representing a first emitted light radiation
from a plurality of wells in a microchip; [0020] b) identifying in
said first image pixels belonging to a boundary line of each of
said plurality of wells; [0021] c) generating a binary image
wherein a first value of light intensity is attributed to pixels
belonging to said boundary lines, and a second value of light
intensity, different from said first value of light intensity, is
attributed to the remaining pixels of said binary image, whereby
said pixels having the first value of light intensity form, in said
binary image, one-dimensional or two-dimensional geometrical
shapes; [0022] d) recognizing among said geometrical shapes, those
shapes that best approximate a reference geometrical shape
identifying an ideal well; [0023] e) generating a second image
wherein said reference geometrical shape is associated with each
geometrical shape defined by the pixels having the first value of
brightness only in the case where said recognizing step has yielded
a positive result, [0024] f) acquiring from said first image,
values of light intensity only at portions of the first image
corresponding to respective portions of the second image internally
delimited the reference geometrical shapes; [0025] g) determining a
result of said biochemical or chemical process in each well on the
basis of said values of light intensity detected in each ideal
well; [0026] h) repeating steps a-g at successive instants in time;
and [0027] i) calculating a quantitative biochemical result based
on the results of steps (g) and (h). [0028] A method for carrying
out a biochemical process, comprising the following steps in order
[0029] a) activating a light source to illuminate a plurality of
wells with a excitation light radiation having a first wavelength
(.lamda..sub.E1) such that an emitted light radiation
(.lamda..sub.E2), having a second wavelength, is emitted by said
wells in response to said excitation light radiation; [0030] b)
acquiring a first image representing said emitted light radiation
from said plurality of wells; [0031] c) identifying pixels
belonging to boundary lines of each of said wells, generating a
binary image wherein a first value of light intensity is attributed
to pixels belonging to said boundary lines, and a second value of
light intensity, different from said first value of light
intensity, is attributed to the remaining pixels of said binary
image, said pixels having the first value of light intensity
defining one-dimensional or two-dimensional geometrical shapes in
said binary image; [0032] d) recognizing, among said geometrical
shapes, the ones that best approximate a reference geometrical
shape identifying an ideal well; [0033] e) generating a second
image in which said reference geometrical shape is associated to
each geometrical shape defined by the pixels having the first value
of brightness only in the case where the recognition has yielded a
positive result; [0034] f) acquiring, from the first image, values
of light intensity only at portions of the first image
corresponding to respective portions of the second image inside
respective reference geometrical shapes associated to respective
geometrical shapes formed by the pixels having the first value of
brightness; and [0035] g) determining a result of said biochemical
analysis on the basis of said values of light intensity detected;
[0036] h) repeating steps a-g at successive instants in time so as
to generate a quantitative biochemical result; and [0037] i)
calculating or determining a quantitative biochemical result based
on the results of steps (g) and (h). [0038] A method as herein
described wherein an identifying step is performed using a Canny
edge-detection filter. [0039] A method as herein described, wherein
a recognizing step is performed using a Hough or Radon filter.
[0040] A method as herein described wherein a determining step
comprises calculating an arithmetic average of the values of light
intensity detected, and supplying a result of the evaluation based
on the mean value of intensity calculated. [0041] A method as
herein described wherein said optical process is a PCR analysis.
[0042] A method as herein described, wherein a calculating step
comprises: [0043] tracing a curve of interpolation of the values of
light intensity acquired for each first image at the successive
instants in time; [0044] concluding that an amplification of said
PCR analysis is successful if the curve is a sigmoid, otherwise
concluding an indication that said amplification of said PCR
analysis failed. [0045] A method as herein described wherein a
calculating step comprises: [0046] calculating a value of light
intensity emitted by each well by performing a sum of the values of
light intensity of a number of pixels of each well and a division
by said number of pixels; and [0047] comparing the calculated value
of light intensity with a threshold, thus determining a value of
fluorescence each of said wells. [0048] A method as herein
described further comprising carry out a plurality of PCR thermal
cycles, wherein steps (a)-(g) are carried out during each PCR
thermal cycle. [0049] A method as herein described further
comprising the steps of [0050] supplying a biological specimen to
be analyzed; [0051] extracting DNA from said biological specimen;
[0052] filling said wells with probes, reagents for PCR, primers
and said DNA; [0053] carrying out a plurality of thermal cycles
each comprising a step of heating to denature said DNA; [0054] a
step of annealing said primers to said DNA, a step of extending
said primers to make amplified DNA and a step of hybridization of
said amplified DNA to said probes. [0055] A method as herein
described wherein the identifying step b is performed using a Canny
edge-detection filter. [0056] A method as herein described wherein
the recognizing step d is performed using a Hough or Radon filter.
[0057] A method as herein described wherein the acquiring step f
comprises calculating the arithmetic average of the values of light
intensity detected, and supplying a result of the evaluation based
on the mean value of intensity calculated. [0058] A method as
herein described wherein the identifying step is performed using a
Canny edge-detection filter; the recognizing step is performed
using a Hough or Radon filter; and acquiring step comprises
calculating the arithmetic average of the values of light intensity
detected, and supplying a result of the evaluation based on the
mean value of intensity calculated. [0059] A method as herein
described further comprising the preceding steps of [0060]
supplying a biological specimen to be analyzed; [0061] extracting
DNA from said biological specimen; [0062] filling said wells with
probes, PCR reagents, a pair of primers, and said extracted DNA;
[0063] carrying out a plurality of thermal cycles comprising a
denaturing step, an annealing step, a primer extension step, and a
hybridization step which generates said emitted light radiation.
[0064] A method as herein described wherein the successive instants
in time occur during said hybridization step. [0065] A method as
herein described wherein the identifying step is performed using a
Canny edge-detection filter. [0066] A method as herein described
wherein the recognizing step is performed using a Hough or Radon
filter. [0067] A method as herein described wherein the acquiring
step comprises calculating the arithmetic average of the values of
light intensity detected, and supplying a result of the evaluation
based on the mean value of intensity calculated. [0068] A method as
herein described wherein the identifying step is performed using a
Canny edge-detection filter, the recognizing step is performed
using a Hough or Radon filter and acquiring step comprises
calculating the arithmetic average of the values of light intensity
detected, and supplying a result of the evaluation based on the
mean value of intensity calculated.
DESCRIPTION OF FIGURES
[0069] For a better understanding of the invention, some
embodiments thereof will now be described purely by way of
non-limiting example and with reference to the attached drawings,
wherein:
[0070] FIG. 1 is a top plan view of a cartridge for biochemical
analyses according to an embodiment of the present invention.
[0071] FIG. 2 is a bottom plan view of the cartridge of FIG. 1.
[0072] FIG. 3 is a lateral view of the cartridge of FIG. 1,
sectioned along the plane of trace III-III of FIG. 1.
[0073] FIG. 4A-4D show, in lateral cross-sectional view, steps for
manufacturing the cartridge of FIG. 1-3.
[0074] FIG. 5 is a perspective view of an analyzer for biochemical
analyses.
[0075] FIG. 6 is a lateral view, sectioned along a longitudinal
plane, of the analyzer of FIG. 5.
[0076] FIG. 7 is a flowchart regarding a method for acquiring
images of the cartridge of FIG. 1-3 by the analyzer of FIG. 5-6 in
order to carry out steps of the biochemical process, according to
an embodiment of the present invention.
[0077] FIG. 8 shows an image of the cartridge of FIG. 1, acquired
by the analyzer of FIG. 5.
[0078] FIG. 9 is a flowchart regarding a method for processing the
image of FIG. 8 in order to identify wells belonging to the
cartridge of FIG. 1 represented in the image of FIG. 8.
[0079] FIG. 10 shows the image of FIG. 8 after the processing steps
of FIG. 9.
[0080] FIG. 11 shows the image of FIG. 9 at the end of the
processing steps of FIG. 7.
[0081] FIG. 12 shows an exploded perspective view of a microreactor
for biochemical analyses of a known type.
[0082] FIG. 13 shows a diagnostic method that may be implemented by
a kit including the microreactor of FIG. 12 and the analyzer of
FIGS. 5 and 6.
DETAILED DESCRIPTION
[0083] According to one aspect of the present disclosure, an
optical analyzer is provided, configured to receive a cartridge for
biochemical analyses including a plurality of wells, and carry out
fluorescence analysis of said cartridge, in particular of the
wells. The analyzer comprises a control unit (microprocessor and
memory) configured to: (a) acquire an image associated to the
fluorescence emitted by the cartridge (in particular, by the
wells); (b) identify, in the image, pixels belonging to boundary
lines of the wells by generating a binary image, i.e., an image
represented with just two levels of color, black and white; (c)
recognize, from among the geometrical shapes identified in step
(b), the ones that best approximate (or resemble) a reference
geometrical shape identifying an ideal well (this step is
performed, in particular by a Hough filter); (d) acquire, from the
starting image, values of light intensity (identifying the
fluorescence emitted by the wells) only in portions of the image
itself that in turn correspond to the regions recognized in step
(c); and (e) evaluate a state of advance and/or a result of the
biochemical analysis on the basis of the values of light intensity
acquired in step (d).
[0084] It is pointed out that, whereas step (b) has the purpose of
locating elements (the wells) in the image, step (c) serves to
interpret what the elements located represent. In fact, step (b)
could yield a not completely correct result; i.e., it could
identify as wells defects of the image or undesirable reflections
of the surface of the cartridge not corresponding to wells. Thus,
step (c) has the function of detecting which shapes identified in
step (b) in actual fact correspond to wells of the cartridge.
[0085] According to a further aspect of the present invention, a
cartridge for biochemical analyses is provided including: a
supporting body, having a first face and a second face opposite to
one another in a direction Z orthogonal to the first and second
faces; a plurality of wells (e.g., ninety-six wells), in particular
adapted to contain a solution for biochemical analyses; one or more
biocompatible layers extending in respective wells; an
anti-reflection layer extending over the first face outside said
wells to generate destructive interference; and a layer extending
within said wells to generate constructive interference. The latter
layer and the biocompatible layer may coincide.
[0086] Furthermore, the cartridge may comprise an on-board
containment module coupled to the first face of the supporting body
(e.g., by gluing or clamping and the like) in outer edge regions of
the first face for surrounding the plurality of wells
completely.
[0087] According to a further aspect of the present disclosure,
likewise provided is a kit, or system, for biochemical analyses
comprising the optical analyzer and the cartridge described
previously, which are designed to co-operate for supplying a result
of the biochemical analysis.
[0088] There now follows a more detailed description of embodiments
of the cartridge and of the optical analyzer, with reference to the
respective figures.
[0089] FIG. 1-3 show a disposable cartridge for biochemical
analyses, designated as a whole by the reference number 1. FIG. 1
is a top plan view of the cartridge 1, FIG. 2 is a bottom plan view
of the cartridge 1, and FIG. 3 is a lateral cross-sectional view,
taken along the line of section III-III of FIGS. 1 and 2.
[0090] The cartridge 1 comprises a supporting body 2 including a
plurality of wells 8, an on-board containment module 3, a heater 5,
and a temperature sensor 6, which form a microreactor for
biochemical analyses. In what follows, the terms "cartridge" and
"microreactor" are used interchangeably in so far as all the
elements that make up the cartridge and the microreactor are
obtained in integrated form using e.g., processes of micromachining
on semiconductor material.
[0091] For simplicity, in what follows reference will be made to
cartridges and instruments for the nucleic-acid amplification by
PCR and the analysis of the results of amplification, without this
implying any limitation. What is described hereinafter, in fact,
finds advantageous application also in systems designed for
execution and recognition of the results of different biochemical
or chemical analyses, in addition to the process of amplification
by PCR.
[0092] In one embodiment, the supporting body 2 is a chip of
semiconductor material, for example monocrystalline silicon, and
has a substantially rectangular shape. Furthermore, a face 2a of
the supporting body 2 presents the plurality of wells 8, and is
coated with a passivation layer 7 of biocompatible material, for
example silicon oxide, BSA, sonicated salmon sperm DNA, and the
like.
[0093] The on-board containment module 3 has e.g., in top plan view
of FIG. 1, a closed polygonal shape and is made of preferably
transparent polymeric material. The on-board containment module 3
is likewise fixed on the passivation layer 7 of the supporting body
2 in peripheral regions of the supporting body 2, for example by
gluing using a biocompatible glue, in such a way as to surround the
wells 8 completely without being superimposed thereon.
[0094] In detail, the cartridge 1 is used in a system for
fluorescence analysis of biological or chemical specimens and is
designed to house the aforesaid specimens. In this context, the
transparency of the on-board containment module 3 is with respect
to the excitation wavelengths and to the wavelengths generated by
the fluorophores present in the specimen contained in the cartridge
1.
[0095] The wells 8 are designed to receive respective specimens of
biological or chemical material for analysis. In one embodiment,
the wells 8 number ninety-six (only some of which are illustrated
in FIG. 1 for simplicity) and are arranged to form a matrix array
having rows and columns. Each well 8 has e.g., a quadrangular
shape, for example square with a side of 800 .mu.m in length and a
depth, measured in the direction Z, of 400 .mu.m. Furthermore, each
well 8 is separated from another well 8 immediately subsequent to
it (along a same row and/or column of the array of wells) by a
distance of between 200 and 500 .mu.m. The wells could also be
cylindrical, or any other suitable shape.
[0096] In general, each well 8 has a capacity between 100 nl and
300 nl, for example equal to about 200 nl. Due to the reduced
dimensions of wells 8, and therefore of the reduced quantity of
samples that each of them contains, the cartridge 1 is preferably
made hydrophilic within the wells 8 and provided with an
antireflective layer out of wells 8. Preferably, the inner walls
and bottom of the wells 8 is covered with a layer adapted to
generate constructive interference of a light beam, which
illuminates the inner portion of the wells 8 (e.g., during
fluorescence imaging).
[0097] In one embodiment, further, the cartridge 1 has been further
functionalized in order to improve biocompatibility by fixing
probes or protein or other ligand to the walls and/or to the bottom
of the wells 8. The probes comprise, for example, single DNA
strands complementary to target sequences being studied in the
biological specimen. The protein can be, e.g., bovine serum albumin
or BSA or antibodies.
[0098] The heater 5 and the temperature sensor 6 are made on a face
2b of the supporting body 2 opposite to the face 2a. In particular,
the heater 5 includes a plurality of resistive coils designed to
develop heat by the Joule effect when they are traversed by
current. The coils of the heater 5 extend over the face 2b in
regions of the face 2b substantially corresponding to respective
regions of the face 2a that house the wells 8. The heater 5 and the
temperature sensor 6 are thermally coupled to the wells 8 so that
the thermal energy released by the heater 5 causes heating of the
biological material in the wells 8. The heater 5 is defined by one
or more conductive paths, for example, of metal or polysilicon. The
temperature sensor 6 is of a thermoresistive type. As is known, in
a thermoresistive sensor, the resistance varies as a function of
the temperature, and thus a reading of the resistance indicates the
temperature at a given moment in time.
[0099] The supporting body 2 further comprises contact pads 9a set
at a longitudinal end of the supporting body 2 to form a connector
9. When the on-board containment module 3 is set on the face 2a of
the supporting body 2, the connector 9 projects to one side with
respect to the on-board containment module 3, outside the area
enclosed by the on-board containment module 3. The connector 9 is
electrically coupled to the heater 5 and to the temperature sensor
6 by conductive paths made of the supporting body 2. The connector
9 enables control of the supporting body 2 (e.g., for carrying out
the thermal PCR cycles) once the supporting body 2 has been
inserted into an analyzer (described hereinafter with reference to
FIGS. 5 and 6).
[0100] To analyze a specimen with the cartridge 1, a mixture of
reagents in solution that comprises fluorophores of two types is
introduced into the wells 8. For instance, a first type of
fluorophores (e.g., FAM fluorophores) has an excitation wavelength
.lamda.E1 and a detection wavelength (or emission wavelength)
.lamda.D1 and combines with a first substance being studied. A
second type of fluorophores (e.g., ROX fluorophores) has an
excitation wavelength .lamda.E2 and a detection wavelength (or
emission wavelength) .lamda.D2 and combines with a second substance
being studied or a control substance. The second type of
fluorophores may have just the function of being a control marker,
whereas the function of molecular probes for detecting the
amplification of DNA is guaranteed by the fluorophores of the first
type.
[0101] According to one aspect of the present disclosure, as
illustrated in FIG. 3 in cross-sectional view, the passivation
layer 7 has a thickness that varies according to whether it is
measured in the wells 8 or else outside the wells 8. In greater
detail, the passivation layer 7 has a thickness outside the wells
8, chosen to function as an anti-reflection layer for the
excitation radiation used during the steps of detection of
fluorescence. In this way, the surface of the cartridge 1 does not
generate undesirable reflections. This problem is not posed inside
the wells 8. On the contrary, within the wells 8 a constructive
interference is preferably generated, to increase the intensity of
the signal emitted by the wells 8. Thus, the portions of the
passivation layer 7 external to the wells 8 are chosen with a
thickness different from the portions of passivation layer 7 inside
the wells 8.
[0102] In practice, to obtain a passivation layer 7 designed to
function as anti-reflection layer outside the wells 8, it is
expedient to consider the wavelength .lamda..sub.E1 of the
excitation radiation, the refractive index n of the material of
which the passivation layer 7 itself is made (in the case where it
is of silicon oxide (SiO.sub.2), this value is approximately
n=1.45), and the cosine of the angle .theta. that is formed with
the normal to the plane of incidence of the radiation. The
effective thickness h of the passivation layer 7, outside the wells
8, is thus chosen between .lamda..sub.E1 cos .theta./(4n)-10% and
.lamda..sub.E1 cos .theta./(4n)+10%, preferably .lamda..sub.E1 cos
.theta./(4n).
[0103] The thickness h of the passivation layer 7, inside the wells
8, is chosen different from .lamda..sub.E1/(4n), for example less
than .lamda.E1/(4n).
[0104] Furthermore, it may be noted that the passivation layer 7 of
silicon oxide (treated e.g., with BSA) is biocompatible and thus
suited to the use described.
[0105] According to an embodiment provided by way of example, the
passivation layer 7 is made of SiO.sub.2 and has a thickness inside
the wells 8 of approximately 20 nm, and alongside the wells 8
(i.e., on the face 2a) of approximately 90 nm.
[0106] FIG. 4A-4D show schematically steps for manufacturing a
cartridge 1 of the type described previously.
[0107] With reference to FIG. 4A, the supporting body 2 is provided
(for example, a silicon substrate in the form of a wafer), having
the face 2a opposite to the face 2b in the direction Z. On the face
2b, the supporting body 2 has the heater 5, the temperature sensor
6, and the connector 9, shown as contact pad 9a. These elements are
formed by known steps, and thus their respective manufacturing
steps are not described in detail.
[0108] Formed on the face 2a is a hard-mask layer 27, of silicon
oxide (SiO.sub.2), for example by thermal growth.
[0109] Then (FIG. 4B), there follows, in a known way and thus not
described in detail, a step of masked etching of the hard-mask
layer 27, for making openings 28 in the hard-mask layer 27 in the
regions of the supporting body 2 in which it is desired to form the
wells 8. The openings 28 expose respective surface portions of the
face 2a.
[0110] Next (FIG. 4C), a further etch, using the hard-mask layer 27
as etching mask, enables removal of selective portions of the
supporting body 2 at the openings 28. The plurality of wells 8 is
thus formed (although in this step they are still lacking the layer
of biocompatible silicon oxide). Etching of the silicon proceeds
until the desired depth for the wells 8 is reached, for example 400
.mu.m.
[0111] Finally (FIG. 4D), a step of thermal growth of silicon oxide
(SiO.sub.2) enables growth of a biocompatible layer 29 inside the
wells 8 and on top of hard-mask layer 27. As an alternative to
thermal growth, it is likewise possible to deposit silicon oxide,
e.g., by chemical vapor deposition (CVD), molecular beam epitaxy
(MBE), Solid Phase Epitaxy (SPE), Liquid Phase Epitaxy (LPE), and
the like.
[0112] The thickness of the hard-mask layer 27 (in particular,
after the step of further growth, or deposition, of FIG. 4D) is
such as for being equal to (.lamda..sub.E1cos .theta.)/(4n), as
illustrated previously, or, more precisely an odd multiple of
(.lamda..sub.E1cos .theta.)/(4n), to prevent undesirable
reflections according to what has been described previously.
Instead, the thickness of the biocompatible layer 29 is different
from (.lamda..sub.E1cos .theta.)/(4n) and more precisely less than
(.lamda..sub.E1cos .theta.)/(4n).
[0113] The hard-mask layer 27 (as obtained after the etching step
of FIG. 4B) together with the biocompatible layer 29 form the
passivation layer 7 illustrated previously.
[0114] After step 4D, a step of gluing of the on-board containment
module 3 is carried out using a biocompatible adhesive, for example
silicone, to obtain the cartridge 1 of FIG. 3, although other
attachment means such as rivets, clamps, welding and the like are
possible.
[0115] In order to use the cartridge 1 in a real-time PCR analyzer,
the wells 8 are further treated in order to improve the
biocompatibility thereof.
[0116] The treatment comprises a step of cleaning and activation,
and also includes a treatment using a solution of CH.sub.3OH:HCl
(4:1) for 10 minutes at room temperature, followed by a step of
rinsing with ultra-pure water at pH 7.0 for removing the reagents
in excess. This is followed by an anhydrification step comprising a
thermal treatment in an oven for 15 minutes at 70.degree. C.
[0117] These steps have the function of rendering the cartridge 1
(in particular, the wells 8) hydrophilic.
[0118] Then, the active surface of the wells 8 is further treated,
during a blocking step, comprising a treatment using a solution
including 1% BSA (bovine serum albumin), 5% SSC (sodium chloride
plus sodium citrate). This step is performed, in particular, at
55.degree. C. for a time of between 4 and 15 hours, where the
solution is left to "rest" in the wells 8.
[0119] Finally, washing is carried out with deionized water.
[0120] Since the treatment that has been described previously to
increase the hydrophilic nature may inhibit PCR on account of the
presence of poly-electrolytes, the latter steps have the function
of restoring characteristics suitable for PCR so that it may take
place correctly and as desired.
[0121] Other surface treatments are possible, as needed for the
application in question.
[0122] FIG. 5 shows a real-time PCR analyzer, designated as a whole
by the reference number 10, designed for being used for reading the
cartridge 1 described previously in order to acquire information on
the state of the PCR.
[0123] As illustrated in FIGS. 5 and 6, the PCR analyzer 10
comprises a first shell 12, closed underneath by a metal plate 13,
and a second shell 14, hinged to the first shell 12. The first
shell 12, the metal plate 13, and the second shell 14 define a
casing of the analyzer 10.
[0124] With reference also to FIG. 6, the first shell 12 has a slot
15 for receiving the cartridge 1. The slot 15 is accessible from
outside for insertion of the cartridge 1 when the second shell 14
is open, in a raised position. In the region enclosed by the
on-board containment module 3 inserted into the slot 15 (i.e., in a
position corresponding to the wells 8), the first shell 12 has a
first window 16 and a second window 17. The first window 16 sets
the slot 15 in communication with the inside of the first shell 12,
whereas the second window 17 enables observation of the wells 8
when the cartridge 1 is inserted into the slot 15 and the second
shell 14 is raised.
[0125] Housed within the first shell 12 are a control board 20, a
fan 21, a collector 22, and a sensor board 23, on which a
calibrated temperature sensor 24 is mounted. The control board 20
and the fan 21 are fixed to the metal plate 13. The control board
20 houses a control unit 25, which presides over operation of the
analyzer 10, as explained hereinafter, and at least one memory
module 26.
[0126] In the embodiment described herein, the fan 21 is aligned to
the windows 16, 17 and may be actuated for drawing in air through
the collector 22. More precisely, a flow of air is drawn in along a
path that develops from the slot 15 to the fan 21 through the
collector 22 in such a way as to cause a heat exchange between the
flow of air and the cartridge 1 set in the slot 15.
[0127] The second shell 14 is hinged to the first shell 12 and
defines a lid, shaped for coupling in a light-tight way with the
first shell 12 and obscuring the second window 17. In practice,
when the second shell 14 is closed on the first shell 12, the
inside of the second shell 14 is substantially inaccessible to
environmental light, and the cartridge 1 inserted into the slot 15
is also obscured. When the second shell 14 is raised, the slot 15
is accessible for inserting and removing the cartridge 1. When the
cartridge 1 is located in the slot 15, further, the wells 8 are
visible and accessible from outside through window 17 to enable
operations of loading of biological specimens for analysis.
[0128] Housed in the second shell 14 are a first light source 30, a
second light source 31 (not visible in cross section of FIG. 6, but
see FIG. 5), a first image sensor 32, and a second image sensor 33
(not visible in cross section of FIG. 6, but see FIG. 5), all
operatively coupled to, and controlled by, the control unit 25.
[0129] The first light source 30 and the second light source 31,
which comprise respective emitter devices 30a, 31a, for example of
the LED type, are oriented for illuminating the cartridge 1 through
the second window 17 and are provided, respectively, with a first
excitation filter 35 and a second excitation filter 36 that
intercept the radiation coming from the emitter device 30a and the
emitter device 31a, respectively. The first excitation filter 35
and the second excitation filter 36 have respective excitation
passbands B.sub.E1, B.sub.E2 centered around excitation wavelengths
.lamda..sub.E1, .lamda..sub.E2 Of fluorophores of two different
types.
[0130] For instance, the first excitation filter 35 has a passband
B.sub.E1 centered around an excitation wavelength .lamda..sub.E1 of
494 nm, i.e., compatible with FAM fluorophores, and the second
excitation filter 36 has a passband B.sub.E2 centered around an
excitation wavelength .lamda..sub.E2 of 575 nm, i.e., compatible
with ROX fluorophores.
[0131] The light radiation supplied by the first light source 30
and by the second light source 31 is thus substantially confined,
respectively, in the excitation passband B.sub.E1 and in the
excitation passband B.sub.E2 of the first excitation filter 35 and
of the second excitation filter 36. The excitation passbands
B.sub.E1, B.sub.E2 are further separate and not overlapping.
[0132] The first image sensor 32 and the second image sensor 33,
for example complementary metal-oxide-semiconductor or "CMOS"
sensors, are arranged for receiving the light emitted by the
fluorophores present in the specimen contained in the cartridge 1
and excited by the light coming from the first light source 30 and
the second light source 31. In the embodiment described, in the
case of FAM fluorophores, the wavelength of the radiation emitted
is 516-522 nm (green), whereas, in the case of ROX fluorophores,
the wavelength of the radiation emitted is 602 nm (red).
[0133] According to one embodiment, the first light source 30 and
the first image sensor 32 are aligned along a first axis X,
parallel to the plane of the supporting body 2 when the latter is
located in the slot 15 and rotated through 450 with respect to a
longitudinal axis of the supporting body 2 in the slot 15. The
second light source 31 and the second image sensor 33 are aligned
along a second axis Y, perpendicular to the first axis X and also
rotated through 45.degree..
[0134] The first image sensor 32 and the second image sensor 33 are
provided, respectively, with a first detection filter 37 and a
second detection filter 38. The first detection filter 37 and the
second detection filter 38 have respective detection passbands
B.sub.D1, B.sub.D2 centered around detection wavelengths (or
emission wavelengths) .lamda..sub.D1, .lamda..sub.D2 of the
respective fluorophores. The passbands B.sub.D1, B.sub.D2 of the
first detection filter 37 and the second detection filter 38 are
further separate and not overlapping and exclude, respectively, the
passbands B.sub.E1, B.sub.E2 of the first excitation filter 35 and
of the second excitation filter 36.
[0135] In the embodiment described, further, the first image sensor
32 and the second image sensor 33 are RGB (red, green, blue) CMOS
sensors and each supply three respective signals for the red,
green, and blue channels. In fact, RGB CMOS sensors comprise a
plurality of photodetectors arranged in an array and each provided
with a respective red, green, or blue filter, with the green
elements in a proportion twice that of the red and blue elements
(RGGB), according to the so-called Bayer filter. An RGB sensor thus
supplies three channel signals, or image signals, one for each of
the fundamental colors red, green, and blue, which are then
combined with local-average operators for reconstructing the
original colors of the image acquired. Each image signal thus
represents the same image filtered with a filter corresponding to
one of the fundamental colors. In what follows, the term "image
signals" S.sub.I will be understood as indicating all the channel
signals regarding a same image or portion of image (possibly even a
single pixel).
[0136] The signals supplied by the first image sensor 32 and by the
second image sensor 33 thus contain information on the response of
each type of fluorophore in the bands of the fundamental colors,
when one or the other between the first light source 30 and the
second light source 31 is activated.
[0137] The control unit 25 exploits the image signals S.sub.I and
information preliminarily stored in the memory module 26 to
determine the presence and concentrations (possibly zero) in the
specimen of substances under investigation, to which the
fluorophores are bound.
[0138] The control unit 25 presides over the operation of the
analyzer 10 and controls execution principally of a thermal cycling
to obtain amplification of the nucleic acids present in the
biological material, for example by the PCR technique and a
procedure of optical detection of specific sequences of nucleotides
("target DNA").
[0139] The cartridge 1 is inserted into the slot 15, and a solution
containing a specimen of e.g., biological material and the
ingredients for e.g., the amplification process is introduced into
the wells 8. Among the other ingredients needed for PCR, the
solution comprises nucleotides (G, A, T, C), primers, a
DNA-polymerase enzyme (for example, TAQ-polymerase), Mg.sup.++,
fluorophores and probes containing single strand
oligonucleotides.
[0140] In this step, no particular precision is required for
introduction of the biological specimen and of the ingredients for
the amplification process in so far as the on-board containment
module 3 is sufficient to prevent said solution from flowing out of
the cartridge 1. According to one aspect of the present disclosure,
the solution and the ingredients for the amplification process are
in liquid form and are introduced into the wells 8 with the aid of
a soft brush (e.g., of hydrophobic polycarbonate). This step is
sufficient to fill the wells 8 so that each of them contains
approximately 200 nL of solution. To prevent the solution from
evaporating, mineral oil may be used, which is introduced into the
wells 8 for covering them completely. Also in this case, the
on-board containment module 3 is designed to restrain the mineral
oil, preventing it from coming out of the cartridge 1.
[0141] The control unit 25 drives the heater 5 and the fan 21,
respectively, for supplying and subtracting thermal energy so that
the temperature in the wells 8 varies cyclically according to a
pre-set profile, which enables the reactions of amplification (in
brief, denaturing, annealing, and extension followed by a
hybridization step). If the specimen being analyzed contains
sequences complementary to the probes, during the hybridization
step fluorophores are incorporated in the hybridized strands, which
are rendered optically detectable. Correspondence of the
temperature with the desired profile is verified using the
temperature sensor 6.
[0142] For the detection of hybridized strands, which contain
fluorophores, the control unit 25 uses the procedure described
hereinafter with reference to FIG. 7.
[0143] After a step of turning-on, with possible calibration (not
described in detail), the control unit 25 acquires (step S1) both
the image signal S.sub.I associated to the fluorophores of the
first type (e.g., FAM fluorophores) with a detection wavelength
.lamda..sub.D1 (that respond to the excitation wavelength
.lamda..sub.E1 of the first light source 30) and the image signal
S.sub.I associated to the fluorophores of the second type (e.g.,
ROX fluorophores) with a detection wavelength .lamda..sub.D2 (that
respond principally to the excitation wavelength .lamda..sub.E2 of
the second light source 31).
[0144] The image signals S.sub.I thus obtained represent images
defined by an array of dots (pixels). FIG. 8 shows, by way of
example, an image 40 of a portion of the cartridge 1 in which the
fluorescence radiation emitted by each well 8 is visible. In the
sequel of the present description, reference will be made to the
processing of just one image (e.g., the image associated to the
fluorophores of the first type) provided by way of example. The
same steps are carried out on the second image (e.g., the image
associated to the fluorophores of the second type).
[0145] The brightness per single pixel is proportional to the power
of fluorescence detected by the image sensors through the
respective detection filters.
[0146] Next (step S2), the control unit 25 selects, in the image 40
of FIG. 8, regions of interest (ROIs), eliminating those portions
of image devoid of significant information. In the embodiment
described, in particular, the regions of interest selected
correspond to the wells 8 of the cartridge 1. The step S2 is thus
aimed at automatically recognizing the portions of the image 40
that include a well 8, excluding from processing the remaining
portions that do not include a well 8. The wells 8 are isolated in
the image 40 to identify the boundary thereof. Consequently, this
step comprises processing of the image 40 for carrying out an edge
detection of the elements present in the image 40.
[0147] Known in the state of the art are numerous edge-detection
techniques starting from a generic image. These techniques are used
in order to detect the points of a digital image in which the light
intensity undergoes a variation above a certain threshold. Sharp
changes of light intensity of an image typically identify
significant changes of the physical reality that the image
represents. With reference to the image 40 of FIG. 8, a high light
intensity of the image corresponds to the area delimited by the
wells 8 that emit fluorescent radiation. The edge-detection
operation generates an image containing much less information than
the original image 40, since details that are not relevant for the
purposes of boundary identification are eliminated, conserving,
instead, the information essential for describing the geometrical
characteristics of the wells 8.
[0148] Known edge-detection methods comprise search-based methods
and methods based upon zero-crossing. Search-based methods
recognize the boundaries seeking the maxima and the minima of the
first-order derivative of the image, typically identifying the
direction in which there is the maximum local gradient.
Zero-crossing methods seek the points in which the second-order
derivative crosses the zero value.
[0149] One of the known methods that may be used for the present
disclosure to implement step S2 of FIG. 7 is known as Canny
algorithm, or Canny method (J. F. Canny, "A computational approach
to edge detection, IEEE Trans., Pattern Analysis and Machine
Intelligence, 8(6), 1986, 679-698).
[0150] For completeness of description, there are now described,
with reference to FIG. 9, sub-steps that implement, according to
one aspect of the present disclosure, an edge-detection method
according to step S2 of FIG. 7. The method of FIG. 9 may be
implemented on a computer using a software program.
[0151] With reference to FIG. 9 (step P1), the image 40 acquired
(in digital format) is processed for detecting values of luminance.
Luminance is defined as the photometric measurement of light
intensity per unit area. The choice of the value of unit of area A
for being considered is the fruit of assessments deriving from the
particular case under examination, such as for example the size of
the image and the desired precision.
[0152] According to one aspect of the present disclosure, the image
40 has dimensions of 300.times.200 pixels, and the unit of area
chosen for calculation of the luminance is one pixel; namely, the
value of luminance of the image 40 is calculated for each pixel.
For this purpose, there is acquired, for each i-th pixel, the value
of red R.sub.i (in the range 0-255), of green G.sub.i (in the range
0-255), and of blue B.sub.i (in the range 0-255), and the
respective value of luminance L is obtained by applying the
following equation:
L.sub.v.sub.--.sub.i=0.2126R.sub.i+0.7152G.sub.i+0.0722B.sub.i
[0153] It is pointed out that the values 0 and 255 are,
respectively, the minimum value and the maximum value allowed for
each pixel represented on 8 bits. Other ranges may be used,
however.
[0154] An array having the same size in pixels of the image 40 is
thus generated (e.g., a 300.times.200 matrix or a vector having 60
000 locations), where each field identifies a value of luminance of
the respective pixel.
[0155] Then (step P2), a first processing of the image of FIG. 8 is
carried out, normalizing the value of contrast pixel by pixel. For
this purpose, the values of red, green, and blue for each pixel are
extrapolated from the image 40, as described with reference in the
previous step P1. Thus, for each i-th pixel, there are obtained the
values of red R.sub.i (in the range 0-255), green G.sub.i (in the
range 0-255), and blue B.sub.i (in the range 0-255). For each i-th
pixel, an operation of normalization is carried out, for each color
component, according to the formula:
C.sub.i-((R.sub.i+G.sub.i+B.sub.i)/3)
[0156] where C.sub.i is the respective color component
(C.epsilon.{R,G,B}) with respect to which normalization is carried
out.
[0157] An array having the same size in pixels of the image 40
(e.g., a 300.times.200 matrix, or a vector with 60,000 locations)
is thus obtained, where each field identifies a normalized value of
contrast for that respective pixel. This step is optional and has
the function of reducing the possible presence of noise in the
image.
[0158] Then (step P3), values of gradient of brightness of the
image obtained in step P2 are calculated using a Gaussian filter.
It is thus expedient to specify the dimensions of the Gaussian
filter, which affect directly the result of the operation. As is
known, Gaussian filters of small dimensions (small radius) enable
recognition of clearer boundaries at the expense of the processing
rate, whereas filters of large dimensions (large radius) guarantee
greater rapidity of execution but are indicated for recognizing
wider and fuzzier boundaries.
[0159] The present applicant has found that a good compromise, for
the particular case forming the subject of the present disclosure,
is obtained by choosing a value of Gaussian-kernel radius of 3 and
a value of Gaussian-kernel width 5.
[0160] Within the area defined by the values chosen for the radius
and for the width of the Gaussian function, the variation of
intensity of the light is acquired, pixel by pixel. Following upon
step P3, a map of gradients is obtained that supplies the value of
amplitude of the gradient for each pixel of the image 40 and the
direction of the gradient. The value of the gradient, together with
an array of values of the modulus (or magnitude) of the gradient
are used as input parameters for defining the thresholds for use in
the thresholding process of the subsequent step P4. A value of
local maximum indicates a high probability of presence of a portion
of a boundary sought. However, this indication is not sufficient to
decide whether a given region corresponds to a boundary region of a
well 8 or else to a central portion thereof. The points
corresponding to the local maxima are considered as belonging to a
boundary and will be taken into consideration in the subsequent
processing steps. There is a local maximum in the points where the
derivative of the gradient goes to zero.
[0161] At the end of the step of searching for local maxima, the
resulting image is defined by an array containing values of levels
of grey that represent possible edge pixels of each well 8. It is
thus expedient to carry out a decision step to decide which pixels
effectively represent an edge.
[0162] Then (step P4), identification of the boundaries is carried
out by thresholding with hysteresis. For this purpose, two
thresholds are defined: a lower threshold T1 and an upper threshold
T2, which are compared with the gradient calculated for each pixel.
If the value of the gradient is less than the lower threshold T1,
the pixel is rejected; if the value of the gradient is higher than
the upper threshold T2, the pixel is accepted as part of a
boundary; if the value of the gradient is between the two
thresholds T1 and T2, the pixel is accepted only if it is
contiguous to a point already accepted previously.
[0163] The presence of two thresholds T1 and T2 solve the
difficulty that would arise in defining a single value of gradient
of brightness to discriminate whether a pixel belongs to a
boundary. It is evident that the values of T1 and T2 may be chosen
case by case, on the basis of the image that is being analyzed. By
way of example, the present applicant has found that acceptable
values are T1=100 and T2=190. These values are compared with the
value of gradient of brightness detected in the previous step P3
(i.e., the comparison is carried out for each measurement of
gradient, on the array of pixels). The range of possible values for
the gradient, calculated on the specimen image of FIG. 8, is
between 0 and 255.
[0164] At the end of this step P4, a binary image is obtained,
where each pixel is marked as belonging to a boundary or else not
belonging to a boundary.
[0165] Then (step P5), a step of generation of the binary image
(i.e., represented with just two colors, or just two levels of
brightness) is carried out, where the boundary pixels are
represented in white (or black) and the remaining pixels are
represented in black (or white). There is thus obtained an image of
the type illustrated in FIG. 10 and designated by the reference
number 50. The image 50 is a binary image, i.e., represented on
just two levels of color (black and white), where the information
associated to a dot (pixel) is represented only by its
position.
[0166] To return to the steps of the method of FIG. 7 (step S3), a
pattern-recognition algorithm is applied to the image 50 of FIG.
10. An example of algorithm that may be used is known as the Hough
algorithm.
[0167] In this context, pattern recognition is based upon
maximum-likelihood estimations, and not on an exact correspondence.
In other words, statistical variations are taken into account in so
far as the wells represented in the image 50 do not have an ideal
quadrangular shape. By way of example, it is possible to use the
Hough algorithm, or any other known pattern-recognition algorithm,
for example Radon algorithm (or transform, or filter).
[0168] In what follows, explicit reference will be made to the
Hough transform (also referred to as "Hough filter" or "Hough
algorithm"), without this implying any loss of generality.
[0169] The Hough algorithm is a method, in a per se known manner,
used for identifying shapes defined analytically (lines, circles,
polygons, etc.) within a digital image.
[0170] The Hough algorithm is founded on the assumption that each
dot (pixel) provides a contribution (also referred to as "vote") to
the definition of a space different from that of the image,
referred to as "parameter space" or "accumulator" (in the case of
the image 50, the space is two-dimensional, which may be
represented by an accumulation vector or matrix, or "array").
[0171] The Hough algorithm receives as input the co-ordinates of
the points belonging to the curve present in the image (i.e., of
the pixels identified as belonging to a boundary), and supplies as
output a parametric description of the set of the curves
recognized, belonging to a fixed analytical figure (as has been
said, a line, a polygon, etc.).
[0172] In detail, once the analytical figure that is for being
sought has been defined and parameterized (e.g., a straight line),
for each pixel detected within the image 50 all the curves that
could pass through that pixel are identified (in the case of a
search for a straight line, the set of the possible curves is the
sheaf of straight lines passing through the pixel), and the
corresponding locations of the accumulation array are incremented
accordingly.
[0173] An accumulation function is thus obtained, which is defined
in the parameter space and the maxima of which determine the
parameters that identify the curves identified in the image
space.
[0174] In the case of recognition of lines, the analytical
description adopted is not the parametric form y=mx+b, but, for
computational reasons, a representation in polar co-ordinates of
the type .rho.=xcos .theta.+ysen .theta. is used, where the
parameters for being identified are the pair (.rho., .theta.),
where .rho. is the distance between the straight line and the
origin of the reference system and .theta. is the angle that the
normal to the straight line forms with the positive x
semi-axis.
[0175] The accumulation matrix A.sub.(.rho.,.theta.) represents the
space in polar co-ordinates (.rho.,.theta.). For each pixel in the
space of the starting image 50, all the curves are calculated that
pass through that pixels and the corresponding cells of the
accumulation matrix A.sub.(.rho.,.theta.) are incremented. The
points of maximum of the accumulation function thus obtained
determine the straight lines sought.
[0176] Starting from equalization of one or more straight lines, it
is possible to identify different geometrical shapes or else
identify the points of intersection between the straight lines,
thus identifying a characteristic point of each well 8.
[0177] Further known in the literature is the generalized Hough
transform, whereby, given a geometrical shape for being identified,
if the image analyzed contains instances of the shape for being
identified, the votes accumulate in the positions of the reference
point that corresponds to said instances. As in the classic Hough
transform, the peak of the accumulator is sought, and said peak
represents the instance of the shape sought. The method described
presupposes that the shapes have the same size and orientation, but
further generalizations are known also in the cases of variation of
scale and orientation.
[0178] For instance, with reference to the image 50 of FIG. 10, it
is possible to use, as input to the function that implements the
Hough transform, the image 50 as an image on which to carry out the
search for the geometrical shapes for being identified and an ideal
reference image that represents each well 8. For instance, the
reference image is a square, for example of 14.times.14 pixels in
size. In this way, by applying the generalized Hough transform, an
image 60 of the type illustrated in FIG. 11 is automatically
obtained, where each instance (the wells 8) of the image 50 deemed
comparable and affine (on the basis of processing of the Hough
transform) with the reference image of FIG. 11 has been
identified.
[0179] The Hough transform also yields a pair of values (p.sub.i,
q.sub.i) for each well 8 identified in the image 60, identifying a
pair of co-ordinates belonging to the respective well 8 (for
example, this pair of co-ordinates identifies a respective vertex
of each well 8). In this way, it is possible to know exactly the
position of each well 8 in the image 60 and, consequently, in the
original image 40.
[0180] FIG. 11 shows an image 60 that is obtained starting from the
image 50 of FIG. 10, where the shapes that, on the basis of the
previous steps, have been identified as resembling the ideal
reference image, have been surrounded by a square 62 that is
precisely the aforesaid ideal reference image. Other shapes (e.g.,
the rounded shapes designated by the reference 65), unlike the
ideal reference square 62, have not been identified as wells 8 and
thus will not be taken into consideration during the subsequent
step S4.
[0181] Next (step S4 of FIG. 7), the fluorescence of the wells 8,
as this may be detected from the image 40, is analyzed. This step
is performed by calculating the values of light intensity of the
pixels of the image 40 exclusively at the wells 8 identified during
the previous step S3, i.e., in regions of the image 40
corresponding to respective regions of the image 60 internally
delimited by the ideal reference squares. More in particular, this
step comprises calculating the mean light intensity at the wells
represented in the image 40.
[0182] For this purpose, therefore, the value of light intensity is
calculated pixel by pixel (as has been said, only for the wells
selected), the values thus calculated are added to one another, and
an operation of division by the total number of pixels considered
is carried out (in other words, an operation of arithmetic average
is performed). The light intensity of the fluorescence emitted by
each well of the image 40 is thus given by the sum of the values of
light intensity of the pixels of each well identified, divided by
the number of pixels, and identifies the value of fluorescence
emitted by each well 8 of the cartridge 1 at the instant of
analysis considered.
[0183] Consequently, it is possible to determine the concentrations
of the emitting fluorophores in the specimen under examination, for
example by thresholds.
[0184] The concentrations thus determined are stored in the memory
module 26, possibly processed by the control unit 25, and made
available through an interface (not illustrated), for example a USB
interface.
[0185] By carrying out the steps S1-S4 of FIG. 7 a number of times
during the PCR thermal cycles it is possible to glean information
on the plot of the fluorescence signal and, consequently,
information on the PCR itself.
[0186] It is known that, in a typical PCR, the PCR product
increases at each amplification cycle, and the diagram of the
fluorescence over the number of cycles exhibits a sigmoidal plot.
In the final cycles, the PCR products no longer increase, and the
curve presents a plateau. Thus, by tracing a curve of interpolation
of the mean-fluorescence values obtained by repeating the steps of
FIG. 7 a number of times it is possible to acquire information
regarding the success of amplification (the curve is substantially
a sigmoid) or otherwise to obtain an indication that the
amplification procedure has not been successful.
[0187] The same steps of FIG. 7 are likewise carried out for
acquiring the fluorescence emitted by the reference fluorophores
(ROX fluorophores) in order to carry out a check with reference
fluorophores.
[0188] The passivation layer 7 with variable thickness enables an
effective reduction of the reflection of the excitation radiation
at the level of the supporting body 2, preventing or reducing
considerably any imprecision in reading due to undesirable
reflections. The advantage is particularly important for portable
analyzers, which, in order for being readily transportable and
usable even outside the laboratory, must be reduced in dimensions
and weight, as well as presenting a low cost. In particular, owing
to the constraints imposed by the applications it is difficult and
economically not advantageous to adopt solutions that might allow
optimization of the uniformity of the radiation detected at the
level of the analyzer.
[0189] According to a further aspect of the present disclosure,
there now follows a description of a procedure for in-vitro
diagnosis of genetic illnesses (FIG. 13) starting from biological
specimens directly taken from the subject undergoing examination
(in particular, specimens of saliva taken with a swabs), directly
automated, in a single step.
[0190] The procedure of FIG. 13 applies, in a particular way, to a
cartridge or microreactor for biochemical analyses of the type
described in the document No. US20130004954, filed in the name of
STMicroelectronics s.r.l., and illustrated in FIG. 12.
[0191] Alternatively, it is also possible to use a cartridge or
microreactor for biochemical analyses of the type described in the
document No. US20130004952, filed in the name of STMicroelectronics
s.r.l.
[0192] However, other cartridges or microreactors may be used,
preferably including wells of a square or rectangular shape having,
in top view, sides of dimensions equal to or greater than 1 mm.
Alternatively, the wells may have a circular shape in top view,
with a diameter equal to or greater than 1 mm.
[0193] The exploded view of FIG. 12 shows a cartridge, or
microreactor, 100 for biochemical analyses of the type described in
US20130004954. The microreactor 100 is housed on a printed-circuit
board (PCB) 102. More precisely, the PCB 102 has a through opening
102a, where the microreactor 100 is housed. The microreactor 100
comprises a first chip 103, for example of polymeric material, and
a second chip 104, of semiconductor material (for example, silicon)
joined together, for example by a silicone-based adhesive.
[0194] A plurality of wells 105 are made in the first chip 103 and
are configured to receiving solutions containing biological
specimens for being analyzed. In one embodiment, the microreactor
100 has been functionalized by fixing DNA probes to the walls or
bottom of the wells 105. The DNA probes may comprise single DNA
strands containing sequences complementary to the target
nucleotides of interest in the biological specimen.
[0195] Integrated in the second chip 104 are heaters 106 and
on-board temperature sensors 107. The on-board temperature sensors
107 are of a thermoresistive type. In practice, their resistance
varies as a function of temperature and thus a reading of the
resistance indicates the temperature at any given instant. The
second chip 104 projects slightly on one side with respect to the
first chip 103 and on the projecting part houses contact pads 108
for connection of the heaters 106 and of the on-board temperature
sensors 107 with conductive paths 109 on the PCB 102. Terminals
109a of the paths 109 enable connection of the PCB 102 once it has
been inserted into a PCR analyzer.
[0196] The microreactor 100 comprises, according to one embodiment,
six wells 105 having a substantially square or rectangular shape,
in top plan view. Each well 105 has, for example, sides of between
3 mm and 4 mm in length, and a depth of approximately 3 mm.
Furthermore, each well 105 is separated laterally from another well
105 by a distance of approximately 1 mm. Of course, 6 wells are
shown for simplicity, but the device can contain additional wells,
such as 12, 24, 48 or 96 wells.
[0197] With reference to FIG. 13, according to the procedure of
diagnosis of genetic illnesses of the present disclosure, a first
step D1 is carried out in which a specimen of buccal swab is
provided. The swab specimen may be acquired by simple sampling of
cells of the buccal epithelium, following a non-invasive procedure,
then dispersing (step D2) the epithelial cells in a transport
medium of a known type (dissolving step).
[0198] The dissolving step dissolution of the specimen cells taken
in a transport medium (or transport solution). The transport
solution is designed to favor lysis of the epithelial cells taken
for releasing the DNA. Lysis is completed during the initial steps
of the thermal cycling of the subsequent steps (favored by
heat).
[0199] For this analysis, each well 105 of the microreactor 100 is
pre-loaded (step D3) with: [0200] 27 .mu.l of wax (which is melted
and poured into each well 8), and this serves to seal each well 8
to prevent evaporation during the PCR thermal cycling; [0201] 4
.mu.l of reagents for PCR (including primers, probes, and
polymerase enzyme), in the form of a liquid solution; and [0202] 2
.mu.l of solution containing the epithelial cells dissolved in the
transport medium.
[0203] The wax used may be compatible with PCR and with the
reagents used; in particular, it must not inhibit PCR. Furthermore,
when molten, it may be transparent and exhibit a low fluorescence
at the wavelengths of interest (used for fluorescence analysis) in
order not to interfere with the measurements of the fluorescence
emitted by the wells 8. Furthermore, it must preferably present a
low vapor tension so as not to evaporate during the thermal cycles
required for PCR.
[0204] Preferably, it must have a density lower than the density of
the PCR reagents and of the solutions used in such a way that it
will cover the PCR reagents and the solution in order to prevent
their evaporation during the PCR thermal cycles. This further
enables introduction into the wells 8 of PCR reagents/solution for
analysis independently before or after insertion of the wax into
the wells 8 (the wells 8 may thus be pre-loaded with the wax during
the steps for manufacturing the cartridge 1 themselves).
Furthermore, the wax enables an effective and valid protection of
the wells 8, thus preventing phenomena of cross-contamination
between adjacent wells 8 and from external contaminating agents.
Furthermore, the wax used must have an adequate melting temperature
so that it is solid at a room temperature (or temperatures lower
than room temperature), but liquid at the temperatures at which PCR
is conducted (typically equal to or higher than approximately
55.degree. C.).
[0205] Applicants have found that a wax that possesses the above
characteristics is a paraffin wax, in particular of the type
marketed by Sigma-Aldrich with the code 76228. As an alternatively
to wax, it is also possible to use mineral oil.
[0206] Then (step D4), the microreactor 100 thus loaded is inserted
into the analyzer (e.g., the analyzer 10 of FIG. 5 or the analyzer
described in the document No. US 2013/0004954), and the PCR thermal
cycling is started.
[0207] For this purpose, by way of example, the thermal cycling
comprises heating to a temperature chosen in the range of 94 to
99.degree. C. for a time of between 2 and 10 minutes, and then a
plurality of cycles (e.g., fifty cycles), where each cycle
includes: [0208] a step of heating to a temperature of between 94
and 99.degree. C. for a time of between 5 and 20 seconds; and
[0209] a step of heating to a temperature of between 57 and
62.degree. C. for a time of between 35 and 70 seconds.
[0210] During these thermal cycles, the PCR process is monitored
(step D5) by fluorescence analysis, for example according to what
has been described previously, by carrying out the steps S1-S4 of
FIG. 7. However, any other real-time monitoring method of a type
known to the state of the art may be used (for example, as
described in the document No. US20130004954).
[0211] Modifications and variations may be made to the device and
to the method described herein, without thereby departing from the
scope of the present invention, as defined in the attached
claims.
* * * * *