Gene identification method, device and system

Abstract

To provide a gene identification method capable of identifying a kind of each gene, even though multiple kinds of genes are present in a same zone.SOLUTION: A gene identification method comprises: a step for combining a melting temperature curve related to multiple kinds of genes, for forming a composite melting temperature curve; a step for using the composite melting temperature curve, for learning a gene identification model 123d set so as to output candidates of multiple kinds of genes in a minute region, at values including an intermediate value; a step for amplifying by PCR (polymerase chain reaction) an identification target gene in a solution distributed in multiple minute regions, then, from a digital PCR measurement unit for measuring fluorescent intensity of the solution distributed in the multiple minute regions, acquiring measurement data indicating the melting temperature curve for each of the multiple minute regions; and a step for inputting the acquired measurement data to the gene identification model 123d, for identifying the multiple kinds of genes in the minute regions.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present invention relates to a gene identification method, a gene identification device, and a gene identification system relating to digital PCR. [Background technology]

[0002] In digital PCR (polymerase chain reaction), first, DNA polymerase, primers, and fluorescently labeled probes required for PCR are added to a limiting diluted sample to prepare a PCR reaction solution. Then, the PCR reaction solution is distributed into micro-areas such as wells or droplets. At this time, the sample is limiting diluted so that one section contains either one molecule of the target gene or none. Next, the target gene in the micro-area is amplified by PCR. Then, the fluorescence intensity of the micro-area is measured, and the number of micro-areas having a fluorescence intensity exceeding a threshold is counted to quantify the target gene. In this way, by using digital PCR, the presence or absence of amplification of the target gene in the micro-area can be output as a value of 1 or 0.

[0003] As described above, the sample is subjected to limiting dilution so that each compartment contains either one molecule of the target gene or none. However, there is a certain probability that each compartment may contain multiple types of genes.

[0004] Therefore, in Patent Document 1, a micro-area containing two different types of specific target genes in one compartment is identified by a DNA detection method including "a step of dividing a DNA solution containing a fluorescently labeled probe or a DNA intercalator and a target DNA into a plurality of compartments, a step of carrying out a nucleic acid amplification reaction in the compartments, a step of measuring the fluorescence intensity with a change in temperature, a step of calculating the melting temperature from a melting curve of a DNA double strand measured based on the change in the fluorescence intensity with the change in temperature, and a step of calculating the temperature difference between two points having a slope of a predetermined value on the melting curve showing the change in the fluorescence intensity." [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2021-3050 Summary of the Invention [Problem to be solved by the invention]

[0006] When the presence or absence of a target gene in a microregion is output as a value of 1 or 0, it is not possible to detect the presence of genes other than the target gene in the microregion. Therefore, even if multiple types of genes are present in the same compartment, the type of each gene cannot be identified. The inventors of the present application have found a method that is different from the method described in Patent Document 1 and that is capable of identifying the type of each gene even if multiple types of genes are present in the same compartment, and have completed the present invention.

[0007] The present invention has been made in consideration of the above-mentioned problems, and aims to provide a gene identification method, a gene identification device, and a gene identification system that are capable of identifying the type of each gene even if multiple types of genes are present in the same compartment. [Means for solving the problem]

[0008] The gene identification method of the present invention includes creating a composite melting temperature curve by combining melting temperature curves for multiple types of genes, learning a gene identification model that is configured to output multiple gene candidates in a microregion with values ​​including an intermediate value using the composite melting temperature curve, amplifying the genes to be identified in a solution distributed in the multiple microregions by PCR (polymerase chain reaction) and measuring the fluorescence intensity of the solution distributed in the multiple microregions, obtaining measurement data indicating the melting temperature curve for each of the multiple microregions from a digital PCR measurement device, and inputting the obtained measurement data into the gene identification model to identify the types of the multiple genes in the microregions.

[0009] Moreover, the gene identification device of the present invention is a gene identification device capable of communicating with a digital PCR measurement device that amplifies genes to be identified in a solution distributed into a plurality of microregions by PCR (polymerase chain reaction) and measures the fluorescence intensity of the solution distributed into the plurality of microregions, and is equipped with a receiving unit that receives measurement data indicating a melting temperature curve for each of the plurality of microregions from the digital PCR measurement device, a memory unit that stores a gene identification model that is set to output candidates for a plurality of genes in the microregion as values ​​including intermediate values, and a processing unit that creates a composite melting temperature curve by combining melting temperature curves for a plurality of types of genes, learns the gene identification model using the composite melting temperature curve, inputs the measurement data received by the receiving unit into the gene identification model, and identifies the types of the plurality of genes in the microregion.

[0010] The gene identification system of the present invention further comprises: a digital PCR measurement device that amplifies genes to be identified in a solution distributed into a plurality of microregions by PCR (polymerase chain reaction) and measures the fluorescence intensity of the solution distributed into the plurality of microregions; a gene identification device capable of communicating with the digital PCR measurement device, the gene identification device comprising: a receiving unit that receives measurement data indicating a melting temperature curve for each of the plurality of microregions from the digital PCR measurement device; a memory unit that stores a gene identification model set to output candidates for a plurality of genes in the microregion with values ​​including an intermediate value; and a processing unit that combines melting temperature curves for a plurality of types of genes to create a composite melting temperature curve, learns the gene identification model using the composite melting temperature curve, inputs the measurement data received by the receiving unit into the gene identification model, and identifies the types of the plurality of genes in the microregions; and a terminal device capable of communicating with the gene identification device, the terminal device displaying the types of the plurality of genes in the microregions identified by the processing unit. Effect of the Invention

[0011] According to the present invention, even if multiple types of genes exist in the same compartment, the type of each gene can be identified. Other objects and novel features will become apparent from the description of this specification and the accompanying drawings. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a configuration diagram of a gene identification system 100. [Diagram 2] FIG. 1 is a diagram showing the configuration of a gene identifying device 120. [Diagram 3] FIG. 13 is a diagram showing melting temperature curves stored in a melting temperature curve DB 123a. [Figure 4] FIG. 13 is a diagram showing composite melting temperature curves stored in a composite melting temperature curve DB 123b. [Diagram 5] FIG. 13 is a diagram showing measurement conditions of wells stored in the measurement information DB 123c. [Figure 6] FIG. 12 shows a conventional gene identification model and a gene identification model 123d. [Figure 7] 13 is a flow chart showing a method for generating composite melting temperature curves for use in training gene identification model 123d. [Figure 8] FIG. 1 shows how to generate composite melting temperature curves when two genes are not amplified with the same primers. [Figure 9] FIG. 1 shows how to generate a composite melting temperature curve when two genes are amplified with the same primers. [Figure 10] 13 is a flowchart showing a learning method of the gene identification model 123d. [Figure 11A] 1 is a graph showing melting temperature curves for genes X to Y. [Figure 11B] 1 is a graph showing melting temperature curves of genes X to Y in which the fluorescence intensities are normalized. [Figure 12] FIG. 1 is a time chart showing a gene identification method. [Figure 13] 13 shows a result display screen 1300 that the gene identifying apparatus 120 displays on the result viewing information terminal 130. FIG. [Figure 14] FIG. 1 is a schematic diagram for explaining problems in the measurement by the dPCR measurement device 110. [Figure 15]FIG. 13 shows learning data used in machine learning of a gene identification model 123d according to the second embodiment. [Figure 16A] 16 shows learning data used to learn a light intensity correction model 1610. FIG. [Figure 16B] FIG. 16 shows learning data used to learn a thermal correction model 1620. [Figure 17] 13 is a flowchart showing a gene identification method using a gene identification model 123d according to the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] In the following embodiments, for convenience, when necessary, the embodiments will be divided into multiple sections or embodiments, but unless otherwise expressly stated, they are not unrelated to each other and one may be a partial or complete variation, detail, supplementary explanation, etc. of the other.

[0014] In addition, in the following embodiments, when referring to the number of elements (including the number, numerical values, amounts, ranges, etc.), unless specifically stated or clearly limited to a specific number in principle, the number is not limited to that specific number and may be more than or less than a specific number.

[0015] Furthermore, in the following embodiments, the components (including element steps, etc.) are not necessarily essential unless specifically stated otherwise or considered to be clearly essential in principle.

[0016] Similarly, in the following embodiments, when referring to the shapes, positional relationships, etc. of components, etc., unless otherwise specified or considered to be obviously different in principle, it includes those that are substantially similar or similar to the shapes, etc. The same applies to the above numerical values ​​and ranges.

[0017] In addition, in all the drawings for explaining the embodiments, the same members are generally given the same reference numerals, and repeated explanations thereof will be omitted.

[0018] (First embodiment) 1 is a configuration diagram of a gene identification system 100. The gene identification system 100 includes a digital PCR measurement device 110, a gene identification device 120, and a result viewing information terminal .

[0019] (Digital PCR measuring device 110) The digital PCR measurement device 110 (hereinafter referred to as the dPCR measurement device 110) divides a PCR reaction solution containing a gene to be identified into many micro-areas, performs PCR on each micro-area, measures the fluorescence intensity of each micro-area, and calculates a melting temperature curve. The digital PCR measurement device 110 amplifies multiple types of genes using multiple primers by multiplex PCR. The measurement data (melting temperature curve) calculated by the dPCR measurement device 110 is transmitted to the gene identification device 120. The dPCR measurement device 110 includes a communication unit 111 that communicates with the gene identification device 120, and a dPCR measurement unit 112 that measures the fluorescence intensity of each micro-area to which the PCR reaction solution is distributed and calculates the melting temperature curve.

[0020] A PCR reaction solution containing DNA polymerase, primers, and fluorescently labeled probes required for PCR added to a specimen (sample) containing a gene of interest that has been subjected to limiting dilution is distributed to micro-areas such as wells or droplets. The dPCR measurement unit 112 then performs a thermal cycle on each micro-area to perform PCR (polymerase chain reaction) amplification. The dPCR measurement unit 112 then measures the fluorescence intensity emitted from the PCR reaction solution for each micro-area and calculates the melting temperature curve for each micro-area.

[0021] The communication unit 111 transmits the measurement data (melting temperature curve) measured by the dPCR measurement unit 112 to the gene identifying device 120. The communication unit 111 may be connected to the gene identifying device 120 via a network, or may be connected locally.

[0022] (Gene Identification Device 120) The gene identification device 120 uses the measurement data received from the dPCR measurement device 110 to identify the type of gene amplified in the microregion. The gene identification device 120 of this embodiment outputs candidates for multiple genes in the microregion with values ​​including intermediate values. For example, when the gene identification device 120 outputs gene A and gene B as candidates for multiple genes in the microregion, the gene identification device 120 outputs gene A: 0.7 and gene B: 0.3 using intermediate values ​​within a predetermined range (for example, 0 to 1). The gene identification device 120 includes a communication unit 121 (receiving unit) that communicates with the dPCR measurement device 110 and the result browsing information terminal 130, a processing unit 122 that performs information processing, and a storage unit 123 that stores various data used in the information processing. The gene identification device 120 may be an on-premise server installed in the same medical institution as the dPCR measurement device 110, or may be a cloud server. The gene identification device 120 will be described in detail later.

[0023] (Result viewing information terminal 130) The result viewing information terminal 130 is a personal computer, smartphone, tablet, or the like for viewing the result data output by the gene identifying device 120. The result viewing information terminal 130 includes a communication unit 131 that communicates with the gene identifying device 120, a control unit 132 that controls the operation of peripheral devices (the communication unit 131, the display unit 133, and the operation unit 134), the display unit 133 that displays various information, and the operation unit 134 that accepts user operations.

[0024] The communication unit 131 may be connected to the gene identifying device 120 via a network or may be connected locally. The communication unit 131 receives result data output by the gene identifying device 120. The display unit 133 is, for example, a liquid crystal display or an organic EL display. The operation unit 134 is, for example, a mouse or a keyboard. The control unit 132 is communicatively connected to the communication unit 131, the display unit 133, and the operation unit 134. The control unit 132 displays the result data received by the communication unit 131 on the display unit 133, for example, in accordance with a user operation accepted by the operation unit 134.

[0025] (Details of gene identification device 120) 2 is a configuration diagram of the gene identifying device 120. The gene identifying device 120 includes the communication unit 121, the processing unit 122, and the storage unit 123 as described above.

[0026] (Communication Department 121) The communication unit 121 is, for example, a network controller that communicates with the dPCR measurement device 110 and the result browsing information terminal 130 via wired or wireless communication.

[0027] (Processing section 122) The processing unit 122 includes a processor 122a, a memory 122b, a communication interface 122c, and an input / output interface 122d. Hereinafter, the interface is abbreviated as I / F. The processor 122a is a central processing unit that performs various calculations. The processor 122a is, for example, a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), etc. The memory 122b stores programs executed by the processor 122a, data required for executing the programs, etc. The memory 122b is, for example, a random access memory (RAM), a flash memory, etc. The communication I / F 122c is an interface that transmits and receives data to and from the communication unit 121, and the input / output I / F 122d is a device controller that inputs and outputs data to and from the storage unit 123, a keyboard, a mouse, a display, etc.

[0028] (Storage unit 123) The storage unit 123 stores various programs and various data. For example, the storage unit 123 stores an OS (Operating System), various programs, various data necessary for executing the programs, etc. For example, the storage unit 123 is a solid state drive (SSD) device, a hard disk drive (HDD) device, or a combination thereof.

[0029] The storage unit 123 includes a melting temperature curve DB 123a, a composite melting curve DB 123b, a measurement information DB 123c, and a gene identification model 123d.

[0030] The melting temperature curve DB123a stores melting temperature curves for a plurality of types of genes. FIG. 3 is a diagram showing melting temperature curves for gene A, gene B, and gene C stored in the melting temperature curve DB123a. The table in FIG. 3 shows the fluorescence intensity (normalized between 0 and 1) at each temperature (40, 50, 60, ... 90) for chip-like wells (#1, 2, 3 ...). Although FIG. 3 shows only melting temperature curves for genes A to C, the melting temperature curve DB123a stores melting temperature curves for the number of types of genes to be identified by the gene identification device 120. The melting temperature curve DB123a stores, in association with each other, identification numbers (#1, 2, 3, ... in the figure) that identify the positions of wells on a chip on which a plurality of wells are formed for use in the measurement by the dPCR measurement device 110 are formed, and melting temperature curves measured in the wells at the positions identified by the identification numbers. The melting temperature curve stored in the melting temperature curve DB123a may be a melting temperature curve calculated from data measured by a dPCR measurement device 110 of a designed PCR reaction solution, or it may be data measured by another dPCR measurement device 110.

[0031] The composite melting temperature curve DB123b stores composite melting temperature curves related to a combination of two genes. FIG. 4 shows a composite melting temperature curve related to a combination of gene A and gene B and a composite melting temperature curve related to a combination of gene A and gene C stored in the composite melting temperature curve DB123b. The table in FIG. 4 shows the total value of the fluorescence intensity at each temperature (40, 50, 60, ... 90) for the chip-like wells (#1, 2, 3 ...). Although FIG. 4 shows only the composite melting temperature curves related to the combination of gene A and gene B and the combination of gene A and gene C, the composite melting temperature curve DB123b stores composite melting temperature curves for the number of combinations of the types of genes to be identified by the gene identification device 120. The composite melting temperature curve DB123b stores, in association with each other, identification numbers (#1, 2, 3, ... in the figure) that identify the positions of wells on a chip on which multiple wells are formed and that are used in measurements by the dPCR measurement device 110, and the composite melting temperature curves of the wells at the positions identified by the identification numbers.

[0032] The measurement information DB123c stores the measurement conditions of the wells in the dPCR measurement device 110. FIG. 5 is a diagram showing the measurement conditions of the wells stored in the measurement information DB123c. The table in FIG. 5 shows the X-coordinates and Y-coordinates of the chip-like wells (#1, 2, 3, ...). For example, as shown in FIG. 5, the measurement information DB123c stores identification numbers (#1, 2, 3, ... in the figure) that identify the positions of the wells on a chip on which multiple wells used in the measurement by the dPCR measurement device 110 are formed, and coordinates (X-coordinates and Y-coordinates) that indicate the positions identified by the identification numbers.

[0033] The gene identification model 123d is a model trained using a composite melting temperature curve and is set to output multiple gene candidates in a well with values ​​including intermediate values. Measurement data measured by the dPCR measurement device 110 is input to the gene identification model 123d. Then, the gene identification model 123d outputs multiple gene candidates in a well with values ​​including intermediate values ​​(e.g., 0.3, 0.7) from the input measurement data.

[0034] 6 is a diagram showing a conventional gene identification model and a gene identification model 123d. The conventional gene identification model outputs the type of gene contained in a well as a value of 1 or 0 based on measurement data showing the melting temperature curve of the well measured by the dPCR measurement device 110. In the example of FIG. 6, the conventional gene identification model outputs the fact that gene C is contained in the well as an identification result "1".

[0035] On the other hand, the gene identification model 123d of this embodiment outputs the type of gene contained in the well as an intermediate value within a predetermined range (0 to 1) from the measurement data indicating the melting temperature curve of the well measured by the dPCR measurement device 110. In the example of Fig. 6, the gene identification model 123d of this embodiment outputs the identification result "0.7" that gene C is contained in the well, and the identification result "0.3" that gene F is contained in the well.

[0036] (Flow of creating a composite melting temperature curve) A method for creating a composite melting temperature curve used for learning the gene identification model 123d will be described. FIG. 7 is a flow chart showing a method for creating a composite melting temperature curve used for learning the gene identification model 123d. Each step in FIG. 7 is executed by the processor 122a of the processing unit 122 of the gene identification device 120 executing a program for creating a composite melting temperature curve (not shown). Note that the composite melting temperature curve may be created by an information processing device different from the gene identification device 120. In this case, the composite melting temperature curve created by the information processing device is stored in the storage unit 123 of the gene identification model 123d.

[0037] The processing unit 122 acquires a melting temperature curve for each type of gene from the dPCR measurement device 110 via the communication unit 121 (S701). Then, the processing unit 122 stores the acquired melting temperature curve for each type of gene in the melting temperature curve DB 123a.

[0038] Next, the processing unit 122 judges whether or not the creation of the composite melting temperature curves related to the combinations of two genes is completed (S702). For example, when creating a composite melting temperature curve related to the combinations of genes A to C, it judges whether or not the creation of the composite melting temperature curve related to the combination of genes A and B, the composite melting temperature curve related to the combination of genes A and C, and the composite melting temperature curve related to the combination of genes A and C is completed.

[0039] When the processing unit 122 determines that the creation of the composite melting temperature curves related to the combinations of two genes is completed (S702: Yes), the processing unit 122 ends this flow. On the other hand, when the processing unit 122 determines that the creation of the composite melting temperature curves related to the combinations of two genes is not completed (S702: No), the processing unit 122 executes the process of S703.

[0040] The processing unit 122 selects two genes related to a combination for which a composite melting temperature curve has not been created, and acquires the melting temperature curves of the selected two genes from the melting temperature curve DB 123a (S703). For example, when a composite melting temperature curve has not been created for a combination of gene A and gene B, the gene identifying device 120 acquires the melting temperature curve related to gene A and the melting temperature curve related to gene B from the melting temperature curve DB 123a.

[0041] Furthermore, the processing unit 122 judges whether or not the two genes selected in S703 are amplified by the same primers (S704).

[0042] If the processing unit 122 determines that the two selected genes are not amplified with the same primer (the two genes are amplified with different primers) (S704: No), it executes the process of S705. On the other hand, if the processing unit 122 determines that the two selected genes are amplified with the same primer (S704: Yes), it executes the process of S706.

[0043] When the processing unit 122 determines that the two selected genes are not amplified with the same primer (the two genes are amplified with different primers) (S704: No), it sums up the values ​​of the melting temperature curves of the two genes for each temperature to create a composite melting temperature curve for the combination (S705). FIG. 8 is a diagram showing a method for creating a composite melting temperature curve when the two genes are not amplified with the same primer. For example, when the two genes A and B selected in S703 are not amplified with the same primer, as shown in FIG. 8, the processing unit 122 sums up the values ​​of the melting temperature curve of gene A and the melting temperature curve of gene B for each temperature to create a composite melting temperature curve for the combination of gene A and gene B.

[0044] On the other hand, when the processing unit 122 determines that the two selected genes are amplified with the same primers (S704: Yes), it creates a composite melting temperature curve by combining the melting temperature curves for the two selected genes in a predetermined ratio (S706, S707).

[0045] For example, the processing unit 122 creates combinations by allocating the amplification ratio of each gene, with the sum of the two selected genes being 1 (S706). For example, the processing unit 122 creates a combination of gene A at 0.9 and gene C at 0.1, a combination of gene A at 0.8 and gene C at 0.2, a combination of gene A at 0.7 and gene C at 0.3, ... a combination of gene A at 0.1 and gene C at 0.9.

[0046] Then, the processing unit 122 multiplies each value of the melting temperature curves of the two genes by the ratio, sums them up, and creates a composite melting temperature curve for the combination (S707). Fig. 9 is a diagram showing a method for creating a composite melting temperature curve when two genes are amplified with the same primer. For example, when the two genes A and C selected in S703 are amplified with the same primer, as shown in Fig. 9, the processing unit 122 creates a composite melting temperature curve for the combination of genes A and C by summing, for each temperature, the values ​​obtained by multiplying each value of the melting temperature curve of gene A by the ratio (for example, 0.7) and the values ​​obtained by multiplying each value of the melting temperature curve of gene C by the ratio (0.3).

[0047] In this manner, the processing unit 122 creates all composite melting temperature curves relating to combinations of two genes.

[0048] (Machine learning flow) A learning method using the above-mentioned composite melting temperature curve will be described. Fig. 10 is a flow chart showing the learning method of the gene identification model 123d. Each step of Fig. 10 is executed by the processor 122a of the processing unit 122 of the gene identification device 120 executing a program for learning the gene identification model 123d.

[0049] The processing unit 122 normalizes the value of each composite melting temperature curve (S1001). The fluorescence intensity of the melting temperature curve and the fluorescence intensity of the composite melting temperature curve in Fig. 8 and Fig. 9 are already normalized values, but if the fluorescence intensity of the melting temperature curve and the fluorescence intensity of the composite melting temperature curve are data before normalization, the fluorescence intensity is normalized at the timing of S1001. The fluorescence intensity normalization may be performed at the time of learning the gene identification model 123d or at the time of creating the composite melting temperature curve.

[0050] Here, normalization of melting temperature curves will be described. Fig. 11A is a graph showing melting temperature curves relating to genes X to Y, and Fig. 11B is a graph showing melting temperature curves obtained by normalizing the fluorescence intensities of the melting temperature curves of genes X to Y. For example, as shown in Fig. 11A, the melting temperature curve acquired from the dPCR measurement device 110 is a data group in which temperature and fluorescence intensity correspond to each other. The gene identification device 120 normalizes the fluorescence intensity of the acquired melting temperature curve to a value between 0 and 1, and creates a normalized melting temperature curve as shown in Fig. 11B.

[0051] Then, the processing unit 122 performs machine learning of the gene identification model 123d using the normalized composite melting temperature curve and gene information related to the normalized composite melting temperature curve to generate a trained gene identification model 123d (S1002). For example, when the normalized composite melting temperature curve is a composite melting temperature curve related to a combination of gene A and gene B, the gene identification model 123d is trained using the normalized composite melting temperature curve, information indicating gene A, information on the amplification ratio of gene A, information indicating gene B, and information on the amplification ratio of gene B as training data.

[0052] (Gene Identification Method) Next, a gene identification method using the trained gene identification model 123d will be described below. Fig. 12 is a time chart showing the gene identification method.

[0053] First, the dPCR measurement device 110 divides a PCR reaction solution containing a sample into multiple wells, performs PCR on each well, measures the fluorescence intensity of each well, and calculates a melting temperature curve (S1201). The dPCR measurement device 110 transmits the measurement data (melting temperature curve) to the gene identification device 120.

[0054] The gene identifying device 120 receives the measurement data (melting temperature curve) (S1202).

[0055] Then, the gene identifying device 120 selects positive wells in which gene amplification is observed using the measurement data (melting temperature curve) (S1203).

[0056] Next, the gene identification device 120 inputs the measurement data (melting temperature curve) of the positive well to the gene identification model 123d, and identifies the gene using the gene identification model 123d (S1204). The measurement data of the positive well input to the gene identification model 123d is normalized in the same manner as the melting temperature curve used for learning. The gene identification device 120 transmits the result data of the gene identification using the gene identification model 123d to the result viewing information terminal 130.

[0057] The result viewing information terminal 130 displays a result display screen based on the received result data (S1205). The result display screen will be described in detail later.

[0058] Furthermore, the gene identification device 120 causes the result viewing information terminal 130 to display a feedback judgment input screen for samples (wells) with low scores for the genes identified in S1204 (S1206). On this feedback judgment input screen, the user can correct the gene identification results for the samples (wells) with low scores. The corrections from the user are used for re-learning the gene identification model 123d (S1207). Note that the gene identification results for samples (wells) that do not have low scores may also be corrected.

[0059] (Result display screen) 13 is a diagram showing a result display screen 1300 displayed by the gene identification device 120 on the result viewing information terminal 130. The result display screen 1300 has a gene identification score display section 1310, a well position display section 1320, and a melting temperature curve display section 1330.

[0060] The gene identification score display unit 1310 displays the number of wells determined to contain only one gene and the number of wells determined to contain one gene and another gene different from the one gene (the number of wells showing the intermediate value). In the example of Fig. 13, the number of wells (count number) that were positive for gene B was 25, and the number of wells (count number) that showed the intermediate value for gene B was 1.

[0061] The well position display section 1320 displays, in a distinguishable manner, the positions of wells that were positive for the gene selected in the gene identification score display section 1310 and the positions of wells that showed an intermediate value. In the example of Fig. 13, gene B is selected in the gene identification score display section 1310, and the well position display section 1320 displays, in a distinguishable manner, the positions of 25 wells that were positive for gene B and the position of one well that showed the intermediate value for gene B.

[0062] The melting temperature curve display section 1330 displays a graph 1331 of the measurement data (melting temperature curve) obtained from the dPCR measurement device 110, melting temperature curves 1332 (blue) and 1333 (green) of two genes calculated from the measurement data, and details 1334 of each melting temperature curve.

[0063] (Gene Identification Method) Here, the gene identification method of the first embodiment will be described. The gene identification method is Combining melting temperature curves for multiple types of genes to create a composite melting temperature curve (S701 to S707 in FIG. 7); A gene identification model 123d is trained so as to output a plurality of gene candidates in a micro-region (well) with values ​​including intermediate values ​​using a composite melting temperature curve (S1001 and S1002 in FIG. 10); A specific target gene in the solution distributed to the multiple micro-regions is amplified by PCR (polymerase chain reaction), and measurement data showing a melting temperature curve for each of the multiple micro-regions is obtained from a digital PCR measurement device 110 that measures the fluorescence intensity of the solution distributed to the multiple micro-regions (S1202 in FIG. 12); The acquired measurement data is input to a gene identification model 123d to identify the types of multiple genes in the micro region (S1204 in FIG. 12).

[0064] In addition, this gene identification method Accepting corrections from the user for the types of the identified genes (S1206 in FIG. 12); and and retraining the gene identification model according to the received corrections (S1207 in FIG. 12).

[0065] Training the gene-specific model 123d as described above includes training the gene-specific model 123d using data that has been normalized from the composite melting temperature curves (FIG. 11B).

[0066] In addition, this gene identification method The method further includes displaying the count number of micro-areas determined to contain only one gene, and the count number of micro-areas determined to contain both the one gene and another gene different from the one gene (gene identification score display section 1310 in FIG. 13).

[0067] Creating the above-mentioned composite melting temperature curve includes creating a composite melting temperature curve by combining melting temperature curves related to a plurality of types of genes in a predetermined ratio (S706 and S707 in FIG. 7).

[0068] (Effects of the first embodiment) By using the gene identification model 123d, which is configured to output multiple gene candidates in a microregion with values ​​including an intermediate value and which is trained using a composite melting temperature curve, it is possible to identify the type of each gene even if multiple types of genes are present in the same section.

[0069] Furthermore, since the gene identification model 123d outputs multiple gene candidates in the micro region as values ​​including intermediate values, even if a gene with a small amount is mixed with other genes in the same section, the gene with a small amount can be identified. Thus, in this embodiment, even if the number of wells in which a gene with a small amount exists is extremely small, it is possible to identify the gene with a small amount by detecting the wells in which the gene exists with other genes.

[0070] Furthermore, the gene identification model 123d is re-learned according to corrections made by the user to the gene identification results, thereby improving the accuracy of identifying the type of gene.

[0071] Furthermore, by using data obtained by normalizing the composite melting temperature curve for learning the gene identification model 123d, the accuracy of identifying the type of gene is improved even if a difference in fluorescence intensity occurs due to the use of a fluorescence filter.

[0072] Furthermore, as shown in FIG. 13, by displaying the number of composite wells, it becomes easier to grasp the number of wells in which a particular gene is mixed with other genes.

[0073] Furthermore, by creating a composite melting temperature curve by combining melting temperature curves for multiple types of genes in a predetermined ratio, it is possible to accurately identify the types of multiple genes even if two genes are amplified at different amplification ratios in the same section.

[0074] (Second embodiment: simple learning type) In the second embodiment, not only the composite melting temperature curve but also the measurement conditions of the wells (microregions) are used for learning the gene identification model 123d. In dPCR, the temperature of the chip is changed, and the fluorescence intensity of each well of the chip is measured during the process. However, there are the following problems. FIG. 14 is a schematic diagram for explaining the problems in the measurement of the dPCR measurement device 110.

[0075] 14, the focus of the camera 113 of the dPCR measurement device 110 is set vertically, and there is variation in the optical distance depending on the position of the well 151 of the chip 150, which causes errors in the intensity and color of the fluorescent color. In addition, in dPCR, although the chip 150 is heated and temperature controlled, temperature unevenness occurs depending on the position of the well 151 of the chip 150, so there is a difference between the measured temperature of the well 151 and the fluorescent intensity.

[0076] In the second embodiment, focusing on the fact that the above-mentioned variations in optical distance and temperature unevenness are caused by the positions of the wells 151 of the chip 150, machine learning of the gene identification model 123d is performed using the positions of the wells 151 on the chip 150 as additional parameters.

[0077] Fig. 15 is a diagram showing learning data used in machine learning of the gene identification model 123d of the second embodiment. As shown in Fig. 15, learning data 1500 used in machine learning of the gene identification model 123d of the second embodiment is a combination of a melting temperature curve 1501 for each well and position information 1502 of the wells on the chip. The position information 1502 of the wells on the chip includes the X coordinate and the Y coordinate of the well 151 on the chip 150.

[0078] (Effects of the second embodiment) By using not only the composite melting temperature curve but also the measurement conditions of the well (micro area) for learning the gene identification model 123d, it is possible to obtain the gene identification model 123d that takes into account the variation in the measurement results due to the measurement conditions of the well. This makes it possible to identify genes even if the mutations have similar melting temperatures (for example, a mutation gene with a single base difference). The other effects are similar to those of the first embodiment.

[0079] (Third embodiment: Comprehensive learning type) In the second embodiment, the gene identification model 123d is trained using training data 1500 that is a simple combination of melting temperature curve 1501 for each well and position information 1502 of the wells on the chip. In the third embodiment, a light intensity correction model that performs light intensity correction on the training data and measurement data according to the well position, and a heat correction model that performs heat correction on the training data and measurement data according to the well position are prepared. The gene identification model 123d is trained using training data corrected by the light intensity correction model and the heat correction model.

[0080] Fig. 16A is a diagram showing learning data used for learning light intensity correction model 1610, and Fig. 16B is a diagram showing learning data used for learning heat correction model 1620. As shown in Fig. 16A, light intensity correction model 1610 performs machine learning using well position 1611, filter type 1612, and differential spectroscopic value 1613 between the standard melting curve and the melting temperature curve. Also, heat correction model 1620 performs machine learning using well position 1621 and difference 1622 between control heat and measured heat at the well position, as shown in Fig. 16B.

[0081] The gene identification model 123d performs machine learning using learning data in which each value of the composite melting temperature curve relating to the combination of two genes is corrected by the light intensity correction model 1610 and the heat correction model 1620.

[0082] Fig. 17 is a flowchart showing a gene identification method using the gene identification model 123d of the third embodiment. Each step in Fig. 17 is executed by the processor 122a of the processing unit 122 of the gene identification device 120 executing a program for identifying a gene (not shown).

[0083] The processing unit 122 acquires the measurement data of each well on the chip from the dPCR measurement device 110 via the communication unit 121 (S1701).

[0084] Next, the processing unit 122 judges whether or not gene identification has been completed for all wells on the chip (S1702).

[0085] When the processing unit 122 determines that gene identification for all wells on the chip is complete (S1702: Yes), it ends this flow. On the other hand, when the processing unit 122 determines that gene identification for all wells on the chip is not complete (S1702: No), it executes the process of S1703.

[0086] The processing unit 122 selects wells for which gene identification has not been completed (S1703).

[0087] The processing unit 122 performs light intensity correction of the measurement data using the light intensity correction model 1610 according to the position of the selected well.

[0088] Furthermore, the processing unit 122 performs light intensity correction of the measurement data according to the position of the selected well using the thermal correction model 1620. The order of the corrections in S1704 and S1705 may be reversed.

[0089] Then, the processing unit 122 inputs the measurement data that has been subjected to the light amount correction and heat correction to the gene identification model 123d, and identifies the genes in the wells using the gene identification model 123d (S1706).

[0090] In this manner, the processing unit 122 identifies the genes in all the wells on the chip.

[0091] (Effects of the third embodiment) By using the light amount correction model 1610 and the heat correction model 1620, the gene identification model 123d can be trained using light amount corrected and heat corrected training data. As a result, the type of gene can be accurately identified taking into account the variation in optical distance and the temperature unevenness. The other effects are similar to those of the first embodiment.

[0092] (Modification) The present invention is not limited to the above-described embodiment, and includes various modified examples. For example, the above-described embodiment has been described in detail to clearly explain the present invention, and is not necessarily limited to those including all of the configurations described. In addition, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. In addition, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

[0093] For example, in the first embodiment, the gene identification model 123d performs machine learning using a composite melting temperature curve relating to a combination of two genes as learning data, but machine learning may also be performed using a composite melting temperature curve relating to a combination of three or more genes as learning data.

[0094] For example, in the third embodiment, the gene identification model 123d performs machine learning using learning data in which each value of a composite melting temperature curve relating to a combination of two genes is corrected by the light intensity correction model 1610 and the heat correction model 1620. However, the gene identification model 123d of the present invention may perform machine learning using learning data corrected by either the light intensity correction model 1610 or the heat correction model 1620. [Explanation of symbols]

[0095] 100: Gene identification system 110: Digital PCR measuring device 120: Gene identification device 130: Results viewing information terminal 121: Communications Department 122: Processing section 122a: Processor 122b:Memory 122c: Communication I / F 122d: Input / output I / F 123: Storage section 123a: Melting temperature curve DB 123b: Composite melting temperature curve DB 123c: Measurement information DB 123d: Gene specific model 1300: Result display screen

Claims

1. Combining melting temperature curves relating to multiple types of genes to create a composite melting temperature curve; training a gene identification model configured to output a plurality of gene candidates in a microregion with values ​​including intermediate values ​​using the composite melting temperature curve; amplifying a gene of interest in a solution distributed to a plurality of micro-regions by PCR (polymerase chain reaction), and acquiring measurement data showing a melting temperature curve for each of the plurality of micro-regions from a digital PCR measurement device that measures the fluorescence intensity of the solution distributed to the plurality of micro-regions; inputting the acquired measurement data into the gene identification model and identifying the types of the multiple genes in the microregion; A method for identifying a gene.

2. accepting corrections from a user to the identified types of genes; and and retraining the gene identification model according to the received corrections. The method for identifying a gene according to claim 1 .

3. Training the gene identification model includes training the gene identification model using the composite melting temperature curve and measurement conditions of the microregion. The method for identifying a gene according to claim 1 .

4. training the gene-specific model includes training the gene-specific model using data normalized to the composite melting temperature curve. The method for identifying a gene according to claim 1 .

5. and displaying the count number of microregions determined to contain only one gene, and the count number of microregions determined to contain the one gene and another gene different from the one gene. The method for identifying a gene according to claim 1 .

6. Creating the composite melting temperature curve includes creating a composite melting temperature curve by combining melting temperature curves related to the plurality of types of genes in a predetermined ratio. The method for identifying a gene according to claim 1 .

7. A gene identification device capable of communicating with a digital PCR measurement device that amplifies a gene to be identified in a solution distributed into a plurality of micro-areas by PCR (polymerase chain reaction) and measures a fluorescent intensity of the solution distributed into the plurality of micro-areas, A receiving unit that receives measurement data indicating a melting temperature curve for each of the plurality of micro-regions from the digital PCR measurement device; a storage unit that stores a gene identification model that is set to output a plurality of gene candidates in a micro region as values ​​including an intermediate value; a processing unit that creates a composite melting temperature curve by combining melting temperature curves related to a plurality of types of genes, learns the gene identification model using the composite melting temperature curve, inputs the measurement data received by the receiving unit into the gene identification model, and identifies the types of the plurality of genes in the microregion. A gene identification device characterized by:

8. The processing unit re-learns the gene identification model according to corrections made by a user to the types of the identified genes. The gene identification device according to claim 7 .

9. The processing unit learns the gene identification model using the composite melting temperature curve and the measurement conditions of the microregion. The gene identification device according to claim 7 .

10. The processing unit trains the gene identification model using data obtained by normalizing the composite melting temperature curve. The gene identification device according to claim 7 .

11. The processing unit outputs a count number of microregions determined to contain only one gene, and a count number of microregions determined to contain the one gene and another gene different from the one gene. The gene identification device according to claim 7 .

12. The processing unit learns the gene identification model using a composite melting temperature curve obtained by combining melting temperature curves related to the plurality of types of genes at a predetermined ratio. The gene identification device according to claim 7 .

13. a digital PCR measuring device that amplifies a specific target gene in a solution distributed to a plurality of micro-areas by PCR (polymerase chain reaction) and measures the fluorescence intensity of the solution distributed to the plurality of micro-areas; A gene identification device capable of communicating with the digital PCR measurement device, A receiving unit that receives measurement data indicating a melting temperature curve for each of the plurality of micro-regions from the digital PCR measurement device; a storage unit that stores a gene identification model that is set to output a plurality of gene candidates in a micro region as values ​​including an intermediate value; a processing unit that creates a composite melting temperature curve by combining melting temperature curves related to a plurality of types of genes, learns the gene identification model using the composite melting temperature curve, inputs the measurement data received by the receiving unit into the gene identification model, and identifies the types of the plurality of genes in the microregion; and a terminal device capable of communicating with the gene identification device, the terminal device displaying the types of the multiple genes in the microregion identified by the processing unit; A gene identification system characterized by:

Images