Detection method, detection device, detection system, detection program, and recording medium

Abstract

This detection method involves: a measurement step for measuring expression levels of multiple nucleic acid molecules included in each of multiple specimens by the same measurement process and acquiring measurement data indicating measurement results of the expression levels; a calculation step for calculating a processed value through unique digitalization processing for converting multiple expression levels in the measurement data into a unique value using measurement data of at least some of the multiple specimens from among the measurement data acquired in the measurement step; and a detection step for comparing the processed value obtained through calculation in the calculation step with a reference value to detect an abnormality occurring in a process until the measurement data are acquired.

Description

DETECTION METHOD, DETECTION DEVICE, DETECTION SYSTEM, DETECTION PROGRAM, AND RECORDING MEDIUM

[0001] The present disclosure relates to a detection method, a detection device, a detection system, a detection program, and a recording medium.

[0002] Japanese Patent No. 7021097 discloses a disease prevalence assessment device including a sample data acquisition unit and a prevalence assessment unit. The sample data acquisition unit acquires sample data including expression levels of multiple types of miRNA in a biological sample. The prevalence assessment unit uses a trained model to output prevalence assessment results for the multiple diseases in the multiple body parts for the acquired sample data. The trained model is a trained model that is capable of determining the prevalence of each of the multiple diseases, including multiple malignant diseases or multiple benign diseases, including cases where the subject is affected by multiple diseases, obtained in advance by machine learning using training data including multiple sample data having items for identifying the presence or absence of the multiple diseases in the multiple body parts.

[0003] WO 2021 / 132547 discloses a testing method for testing for a disease using a disease marker, the testing method including a specimen data acquisition step and a discrimination step. The specimen data acquisition step acquires marker data indicating the results of measuring a disease marker in a body fluid sample collected from a subject, and preparation data indicating the preparation conditions of the body fluid sample. The discrimination step determines the presence or absence of a disease in the subject by inputting the marker data and preparation data acquired in the specimen data acquisition step into a trained model that has been machine-learned to determine the correlation between a set of marker data indicating the results of measuring a disease marker in a body fluid sample and preparation data indicating the preparation conditions of the body fluid sample, and the presence or absence of a disease in the subject from whom the body fluid sample was collected.

[0004] When measuring the expression levels of multiple nucleic acid molecules contained in a sample and determining the disease of the sample provider based on measurement data indicating the measurement results, any abnormality occurring in the process leading up to obtaining the measurement data will affect the determination results. Examples of such abnormalities include abnormalities due to improper processing in each step, such as pre-processing of the sample before measurement, measurement processing, and acquisition processing for obtaining the measurement results as measurement data.

[0005] One possible method for detecting such abnormalities is, for example, a method (referred to as the conventional PSI method) that uses a single standard sample to define a reference distribution that serves as the detection standard and detects the abnormalities using an index called PSI (Population Stability Index).

[0006] An example of the use of the conventional PSI method is the measurement of microRNA expression levels using a next-generation sequencer (NGS) over multiple measurements. In this example, for example, serum from the same donor is included in all measurements. In this example, for example, the expression level distribution of a standard sample from the measurement to be evaluated is compared with the expression level distribution of a standard sample from the same donor measured in the past, and an index value (PSI value) is calculated. The obtained PSI value is compared with a predetermined arbitrary threshold, and an abnormality is detected if it exceeds the threshold.

[0007] As described above, detection methods that use standard samples are subject to constant deterioration and fluctuation of the standard sample itself, and it is impossible to permanently obtain the same standard sample, making continuous operation difficult. Furthermore, detection methods that use a single standard sample are susceptible to the effects of deterioration of the standard sample, and detection results tend to vary, which may result in reduced detection accuracy.

[0008] The deterioration of standard samples may be caused by the following three factors: (1) the sample being left at room temperature; (2) a long time having passed since the sample was collected; and (3) the sample being repeatedly frozen and thawed.

[0009] An object of the present disclosure is to provide a detection method, a detection device, a detection system, a detection program, and a recording medium that can detect with high accuracy abnormalities that occur in the process of obtaining measurement data that indicates the measurement results of the expression levels of multiple nucleic acid molecules contained in a sample.

[0010] A detection method according to one aspect of the present disclosure includes a measurement step of measuring the expression levels of multiple nucleic acid molecules contained in each of multiple samples using the same measurement process and obtaining measurement data indicating the measurement results of the expression levels; a calculation step of calculating a processed value using the measurement data for at least some of the multiple samples obtained in the measurement step by a unique digitization process that converts the multiple expression levels in the measurement data into a unique value; and a detection step of comparing the processed value calculated in the calculation step with a reference value to detect any abnormalities that occurred in the process leading up to obtaining the measurement data.

[0011] According to the present disclosure, it is possible to detect with high accuracy abnormalities that occur in the process leading up to obtaining measurement data indicating the measurement results of the expression levels of multiple nucleic acid molecules contained in a sample.

[0012] FIG. 1 is a flow diagram showing each step of a detection method according to the present embodiment. FIG. 2 is an example of an equation used in the unique digitization process according to the present embodiment. FIG. 3 is a schematic block diagram showing an example of a computer that functions as the detection device according to the present embodiment. FIG. 4 is a block diagram showing the functional configuration of the detection device according to the present embodiment. FIG. 5 is a graph showing processed values ​​at each measurement in Example 1-1. FIG. 6 is a graph showing processed values ​​at each measurement in Example 1-2. FIG. 7 is a graph showing processed values ​​at each measurement in Example 1-3. FIG. 8 is a graph showing processed values ​​at each measurement in Example 1-4. FIG. 9 is a graph showing processed values ​​at each measurement in Example 2.

[0013] An example of an embodiment of the technology of the present disclosure will be described below with reference to the drawings. Note that components and processes that perform the same operations, actions, and functions are given the same reference numerals throughout the drawings, and redundant explanations may be omitted as appropriate. Each drawing is merely a schematic illustration to allow a sufficient understanding of the technology of the present disclosure. Therefore, the technology of the present disclosure is not limited to the illustrated examples. Furthermore, in this embodiment, explanations of configurations that are not directly related to the present disclosure or well-known configurations may be omitted.

[0014] <Detection Method 10> First, a description will be given of a detection method 10 according to this embodiment. Fig. 1 is a schematic diagram showing each step of the detection method 10 according to this embodiment.

[0015] Detection method 10 is a method for measuring the expression levels of multiple nucleic acid molecules contained in each of multiple samples using the same measurement process, obtaining measurement data indicating the measurement results of the expression levels, and detecting abnormalities that occur in the process leading up to obtaining the measurement data.

[0016] Examples of specimens that are the measurement targets for measuring the expression levels of multiple nucleic acid molecules include body fluids (e.g., blood, serum, urine, tears, saliva, sweat, semen, lymph, tissue fluid, body cavity fluid (e.g., pleural effusion, ascites, etc.), cerebrospinal fluid, amniotic fluid, vaginal fluid, nasal mucus), tissues, and cells. The specimen may be any specimen that allows measurement of the expression levels of multiple nucleic acid molecules. The specimen may be collected from humans or non-human animals. Examples of non-human animals include non-human mammals (monkeys, dogs, cats, mice, rats, rabbits, cows, horses, pigs, sheep, etc.), birds (chickens, quails, etc.), etc.

[0017] Examples of nucleic acid molecules include microRNA. The nucleic acid molecule may be small RNA other than microRNA, other RNA, DNA, or the like. Furthermore, the nucleic acid molecule does not have to be a nucleic acid molecule composed only of ATGCU bases. Specific examples include nucleic acids that have undergone modifications such as DNA / RNA methylation, or edits such as A-to-I RNA editing. Thus, various nucleic acid molecules can be targets for application of this detection method.

[0018] 1 , the detection method 10 preferably includes a measurement step 12, a calculation step 13, and a detection step 14, and further includes a disease determination step 16. In this embodiment, the measurement step 12, the calculation step 13, the detection step 14, and the disease determination step 16 are performed in this order, for example.

[0019] Since the detection method 10 is a method including a measurement step 12, a calculation step 13, and a disease assessment step 16, it can also be called a measurement method, a calculation method, or a assessment method. Furthermore, when testing or analysis is performed based on the assessment in the disease assessment step 16, the detection method can also be called a testing method or an analysis method. Each step of the detection method 10 will be described below.

[0020] <Measurement Step 12> The measurement step 12 is a step of measuring the expression levels of a plurality of nucleic acid molecules contained in each of a plurality of samples by the same measurement process, and obtaining measurement data indicating the measurement results of the expression levels.

[0021] Specifically, in the measurement step 12, for example, a next-generation sequencer (NGS) is used as a measurement device to measure multiple nucleic acid molecules contained in each of multiple samples and identify the base sequence of each nucleic acid molecule. Next, the number of identified nucleic acid molecules is counted for each base sequence to determine the number of reads of the nucleic acid molecule in the NGS. This number of reads of the nucleic acid molecule corresponds to the expression level (specifically, absolute expression level) of the nucleic acid molecule.

[0022] Then, in the measurement step 12, measurement data indicating the measurement results (specifically, the number of reads) obtained by measuring the expression levels of multiple nucleic acid molecules in a specimen from a subject is obtained. Here, the measurement data is measurement data for determination to be used for determination in the disease determination step 16, and may be data processed by performing data processing on the measurement results. In other words, the measurement data may be data indicating the expression levels of nucleic acid molecules as relative values. An example of such data processing is a conversion process in which the measurement results obtained by measuring the expression levels of multiple nucleic acid molecules are converted into measurement data for determination. The measurement data is obtained, for example, as multidimensional data indicating the expression levels of multiple nucleic acid molecules for each type of nucleic acid molecule.

[0023] The measurement data may be absolute quantitative values ​​in which the expression levels of nucleic acid molecules are quantified as absolute values. Thus, the measurement data showing the measurement results of the expression levels of multiple nucleic acid molecules may be data shown as relative values ​​or data shown as absolute values.

[0024] Furthermore, in the measurement step 12, as described above, the expression levels of nucleic acid molecules are measured using the same measurement process. Here, the same measurement process includes at least measurement using the same measurement device (i.e., a single measurement device). Furthermore, the same measurement process may be considered to mean measurement using the same measurement device under the same measurement conditions.

[0025] In NGS, it is possible to measure the expression levels of multiple nucleic acid molecules in multiple samples at once. Therefore, in the measurement step 12, the expression levels of multiple nucleic acid molecules contained in each of the multiple samples can be measured in a single measurement run using the same measurement device under the same measurement conditions. Therefore, in this embodiment, measurements in the same measurement process can be considered as measurements being performed in a single measurement run using the same measurement device under the same measurement conditions.

[0026] In addition to next-generation sequencers, DNA chips, quantitative PCR, flow cytometers, and the like can also be used as measuring devices as long as they can measure the expression levels of multiple nucleic acid molecules. As described above, if the expression levels of multiple nucleic acid molecules can be measured and nucleic acid molecules that show the measurement results can be obtained, various methods, including known methods, can be used as a method for measuring the expression levels of multiple nucleic acid molecules.

[0027] Furthermore, pre-processing (e.g., centrifugation, storage, library preparation, etc.) may be performed on the sample before the measurement step 12 is performed. In this embodiment, when the pre-processing is performed, the same processing is performed on multiple samples. That is, in this embodiment, pre-processing is performed on multiple samples using the same process.

[0028] The same measurement process means that measurements are taken using at least the same measurement device (i.e., a single measurement device), and the measurement conditions and number of measurements in the measurement device, as well as the pre-processing before the measurement, may or may not be the same.

[0029] <Calculation step 13> The calculation step 13 is a step of calculating a processed value by a unique digitization process using measurement data indicating the expression levels of multiple nucleic acid molecules in at least some of multiple samples, among the measurement data obtained in the measurement step 12. In this embodiment, for example, measurement data (multidimensional data) regarding all expression levels of all samples obtained in the measurement step 12 is used.

[0030] Here, the unique digitization process is a process of converting multiple expression levels in the measurement data into a unique value. That is, the unique digitization process is a process of converting multivariates, namely the expression levels of multiple nucleic acid molecules contained in each of multiple samples, into a unique value. Specifically, in this embodiment, as an example, the unique digitization process is performed using a method using the PSI (Population Stability Index) in the following procedure.

[0031] First, the expression level distribution of the expression levels of multiple nucleic acid molecules in one sample related to the measurement data is compared with a reference distribution to determine a processing value (PSI value) (first processing). Specifically, the processing value (PSI value) is calculated from the distribution deviation between the expression level distribution and the reference distribution using the formula shown in Figure 2.

[0032] The reference distribution can be an expression level distribution based on expression levels measured in an earlier measurement than the measurement related to the expression level distribution. The first process is performed on all samples obtained in the measurement step 12 (i.e., all samples in a single measurement), and the PSI values ​​for all samples are calculated.

[0033] Then, a representative PSI value is calculated from the PSI values ​​of all the samples (second process). Examples of the representative PSI value calculated in the second process include the average value and median of the PSI values ​​of all the samples.

[0034] As described above, in this embodiment, the PSI value is not calculated from the sum of the expression levels of all samples, but rather the PSI value is calculated for each sample, and a representative PSI value is calculated from the PSI values ​​of all samples.

[0035] In the above example, a method using PSI was applied as the unique digitization process, but this is not limited to this. For example, distance functions such as EMD (Earth-Mover distance), information criteria such as KLD (Kullback-Leibler divergence) and JSD (Jensen-Shannon divergence) may also be used, and various statistical methods can be used. The values ​​calculated by the unique digitization process using the methods using EMD, JSD, and KLD are referred to as the EMD value, JSD value, and KLD value, respectively. More specifically, the methods using EMD, JSD, and KLD are used to compare the expression level distribution of the expression levels of multiple nucleic acid molecules in one sample related to the measurement data with a reference distribution to determine the processed values, the EMD value, the JSD value, and the KLD value, respectively. Furthermore, while a unique digitization process using a single method was used, this is not limited to this. For example, it is possible to combine values ​​calculated using different methods, such as by taking the arithmetic mean of a unique value calculated using PSI (PSI value) and a unique value calculated using KLD (KLD value). Examples of methods for combining include arithmetic averaging and weighted averaging.

[0036] <Detection Step 14> The detection step 14 is a step of comparing the processed value calculated in the calculation step 13 with a reference value to detect an abnormality that occurred during the process up to the acquisition of the measurement data.

[0037] The reference value may be, for example, a past processed value calculated before the current processed value. Therefore, in the detection step 14, the processed value calculated in the calculation step 13 is compared with a past processed value calculated before the current processed value as the reference value to detect an abnormality.

[0038] Furthermore, the past processed value may be a value calculated from processed values ​​in a plurality of measurements taken before the measurement at which the processed value was measured. For example, the past plurality of measurements may be selected from a plurality of consecutive measurements taken immediately before the measurement at which the processed value was measured. In this way, when calculating a reference value from a plurality of values, the reference value may be calculated using, for example, the average, median, or quartile.

[0039] In the detection step 14, it is possible to detect an abnormality, for example, based on whether the processed value is higher or lower than the reference value, or the degree of difference between the processed value and the reference value. Furthermore, in the case where a past processed value is used as the reference value in the detection step 14, an abnormality may be detected based on whether the processed value is higher or lower than the reference value multiple times in succession. In other words, it is possible to detect an abnormality based on whether the processed value tends to increase or decrease as a result of comparing the processed value with the past processed value.

[0040] Here, the abnormality to be detected in the detection step 14 is an abnormality that occurs in the process up to acquiring the measurement data. Specifically, the abnormality is an abnormality that occurs in the process from collecting a sample from a subject to acquiring the measurement data. Examples of the abnormality include an abnormality caused by inappropriate processing in each step, such as the collection process for collecting the sample, pre-processing of the sample before measurement (e.g., processing such as transportation, storage, and library preparation), measurement process, and acquisition process for acquiring the measurement results as measurement data.

[0041] The acquisition process may, for example, be a process of converting the measurement results (i.e., raw data) obtained by measuring the expression levels of a plurality of nucleic acid molecules into measurement data for use in disease diagnosis.

[0042] Furthermore, examples of abnormalities to be detected in the detection step 14 include abnormalities caused when, based on measurement data, the properties of each of a sample group containing multiple samples are evaluated, and the sample group to be evaluated is different from the evaluation target assumed in the evaluation by the evaluation algorithm that is the evaluation standard. Specifically, for example, when the property is evaluated as the presence or absence of a disease, there are cases where, in an evaluation algorithm that assumes that the frequency of evaluation for the presence of a disease is low, the sample group to be evaluated contains a large number of samples from diseased patients.

[0043] <Presence determination step 16> The prescribing step 16 is a step of determining the presence or absence of a disease based on the measurement data obtained in the measuring step 12. Specifically, in the prescribing step 16, the presence or absence of a disease is determined by inputting the measurement data into a disease-trained model that has been machine-learned to determine the correlation between data indicating the results of measuring the expression levels of multiple nucleic acid molecules in a sample and prescribing data indicating the presence or absence of a disease in the subject from whom the sample was collected. The measurement data used in the prescribing step 16 is the measurement data obtained in the measuring step 12 and used to calculate the processed value in the calculation step 13.

[0044] In the disease determination step 16, the presence or absence of a disease is determined if no abnormality is detected in the detection step 14. In other words, if an abnormality is detected in the detection step 14, the disease determination step 16 is not executed. In this case, for example, the detection result indicating that an abnormality has been detected may be presented to the user (i.e., the person executing the detection method 10).

[0045] As described above, if an abnormality is detected in the detection step 14, the disease determination step 16 is not executed, but this is not limited to this. For example, even if an abnormality is detected in the detection step 14, the disease determination step 16 may be executed if, for example, a determination result is obtained as reference data. In this case, for example, after a detection result indicating that an abnormality has been detected is obtained, the fact that a determination has been made may be presented to the user. Furthermore, although the disease determination step 16 is executed after the detection step 14 in this embodiment, it may also be executed before the detection step 14. In this case, if an abnormality is detected in the detection step 14, the determination result of the disease determination step 16 is treated as, for example, reference data.

[0046] In the disease determination step 16, the presence or absence of a disease is determined as an evaluation step for evaluating the properties of the sample (in other words, the properties of the subject), but this is not limited to this. For example, the evaluation step may involve basic biological analysis, regression prediction, various abnormality detection, etc., as the properties of the sample based on the measurement data acquired in the measurement step 12. Furthermore, the evaluation step including the disease determination step 16 is an optional step, and the detection method 10 does not necessarily have to include an evaluation step including the disease determination step 16.

[0047] <Detection System 20> Next, a description will be given of the detection system 20 that executes the above-described detection method. The detection system 20 includes a measurement device 21 and a detection device 30, as shown in FIG.

[0048] <Measuring Device 21> The measuring device 21 is an example of a measuring unit, and is a device that executes the above-mentioned measuring step 12. That is, the measuring device 21 measures the expression levels of multiple nucleic acid molecules contained in multiple samples using the same measurement process, and obtains measurement data that indicate the measurement results of the expression levels. As the measuring device 21, for example, an NGS is used.

[0049] The measuring device 21 may be composed of multiple devices. For example, when performing data processing on the measurement results of measuring the expression levels of multiple nucleic acid molecules, the measuring device 21 may be composed of, for example, a device that measures the expression levels of multiple nucleic acid molecules and a processing device that processes the data.

[0050] <Detection Device 30> The detection device 30 is an example of a calculation unit and an example of a detection unit. The detection device 30 is a device that executes the calculation step 13 and the detection step 14 described above. That is, the detection device 30 calculates a processed value by a unique digitization process using measurement data that indicates the expression levels of multiple nucleic acid molecules in at least a portion of multiple samples, among the measurement data acquired by the measurement device 21. Furthermore, the detection device 30 compares the calculated processed value with a reference value to detect abnormalities that occurred during the process of acquiring the measurement data.

[0051] The detection device 30 also executes the aforementioned disease determination step 16. That is, the detection device 30 determines whether or not the subject is suffering from a disease based on the measurement data acquired by the measurement device 21.

[0052] 3, the detection device 30 has a function as a computer, and includes a CPU (Central Processing Unit) 31, a ROM (Read Only Memory) 32, a RAM (Random Access Memory) 33, a storage 34, an input unit 35, a display unit 36, and a communication interface (I / F) 37. The components are connected to each other via a bus 39 so as to be able to communicate with each other.

[0053] The CPU 31 is a central processing unit that executes various programs and controls each part. That is, the CPU 31 reads programs from the ROM 32 or the storage 34 and executes the programs using the RAM 33 as a work area. The CPU 31 controls each of the above components and performs various arithmetic processing in accordance with the programs stored in the ROM 32 or the storage 34. The CPU 31 is an example of a processor.

[0054] The ROM 32 stores various programs and various data. The RAM 33 temporarily stores programs or data as a working area. The storage 34 is configured by an HDD (Hard Disk Drive) or an SSD (Solid State Drive) and stores various programs including the operating system and various data.

[0055] In this embodiment, for example, a detection program for executing a detection process that performs the above-described detection method 10 is recorded in the storage 34. The detection program may be a single program, or a group of programs configured from multiple programs or modules. The detection program may be recorded in the ROM 32. The ROM 32 and the storage 34 function as an example of a non-transitory recording medium.

[0056] An example of a processor is not limited to the above-mentioned CPU, which is a general-purpose processor, but may be, for example, a dedicated processor configured with a circuit designed specifically for executing a specific process. Also, an example of a processor is not limited to a single processor, but may be a processor configured by multiple processors located in physically separate locations working together.

[0057] The input unit 35 includes a pointing device such as a mouse and a keyboard, and is used to input various information. The input unit 35 also receives information on the expression levels of multiple nucleic acid molecules measured by the measurement device 21 as input.

[0058] The display unit 36 ​​is, for example, a liquid crystal display, and displays various information. The detection device 30 can notify the user through the display unit 36 ​​that a target abnormality has been detected.

[0059] The communication interface 37 is an interface for communicating with other devices, and uses standards such as Ethernet (registered trademark), FDDI (Fiber Distributed Data Interface), and Wi-Fi (registered trademark).

[0060] As shown in FIG. 4 , in the detection device 30 , the CPU 31 executes a detection program to function as a calculation unit 150 , a detection function unit 160 , and a determination unit 170 .

[0061] The calculation unit 150 executes the above-mentioned calculation step 13. That is, the calculation unit 150 calculates a processed value by a unique digitization process using measurement data indicating the expression levels of multiple nucleic acid molecules in at least some of multiple samples, among the measurement data acquired by the measurement device 21 (see the above-mentioned calculation step 13).

[0062] The detection function unit 160 executes the above-mentioned detection step 14. That is, the detection function unit 160 compares the processed value calculated by the calculation unit 150 with a reference value to detect an abnormality that occurred during the process up to acquiring the measurement data (see the above-mentioned detection step 14).

[0063] The determination unit 170 executes the disease determination step 16. That is, the determination unit 170 determines the presence or absence of disease based on the measurement data acquired by the measurement device 21 (see the disease determination step 16 described above).

[0064] In the determination device 30, the CPU 31 may perform processing to output the abnormality detection result and the disease presence / absence determination result. This processing is performed, for example, by displaying on the display unit 36 ​​and transmitting to an external device (e.g., the cloud).

[0065] In this embodiment, the detection system 20 includes the measurement device 21 and the detection device 30, but the detection system 20 may be configured with a single device. In this case, the single device functions as an example of a measurement unit, a calculation unit, and a detection unit.

[0066] Furthermore, the detection device 30 may be configured with a plurality of devices. For example, the detection device 30 may be configured with a plurality of devices (e.g., three devices) that share the calculation step 13, the detection step 14, and the disease determination step 16.

[0067] <Effects of this embodiment> In this embodiment, as described above, in the calculation step 13, a processed value is calculated by a unique digitization process that converts multiple expression levels in the measurement data into a unique value using the measurement data for at least a portion of multiple samples from the measurement data acquired in the measurement step 12. Then, in the detection step 14, the processed value calculated in the calculation step 13 is compared with a reference value to detect any abnormalities that occurred in the process up to acquiring the measurement data.

[0068] In this way, in this embodiment, a processed value calculated by a unique digitization process that converts multiple expression levels of measurement data for multiple samples into a unique value is used, so that abnormalities can be detected with high accuracy, while being less susceptible to the effects of sample deterioration and measurement errors compared to when a processed value calculated from measurement data for a single sample is used. Furthermore, in this embodiment, the use of a standard sample is unnecessary, and a decrease in detection sensitivity due to lot variations in standard samples can be avoided compared to conventional PSI methods.

[0069] Specifically, in the detection step 14, the processing value calculated in the calculation step 13 is compared with a past processing value calculated before the processing value in question as a reference value to detect an abnormality.

[0070] In this way, by comparing the past processing value calculated before the current processing value as a reference value, abnormalities can be detected based on the performance of past processing values, thereby reducing variation in the accuracy of abnormality detection.

[0071] Furthermore, in this embodiment, a value obtained from processed values ​​in a plurality of measurements prior to the measurement in which the processed value was measured is used as the past processed value.

[0072] This allows for more valid anomaly detection than when using values ​​derived from processed values ​​from a single measurement run. Specifically, it allows for variations within the specifications of reagents that are not a problem on a daily basis, making it possible to detect only statistically rare events.

[0073] Furthermore, in this embodiment, in the measurement step 12, a processed value is calculated by a unique digitization process using measurement data on all expression levels of all specimens measured by the same measurement process.

[0074] Therefore, compared to using measurement data for only a portion of the samples among the measurement data acquired in the measurement step 12, the method is less susceptible to sample deterioration and measurement errors, and abnormalities can be detected with high accuracy.

[0075] In addition, in this embodiment, in the disease determination step 16, if no abnormality is detected in the detection step 14, it is determined whether the sample provider is suffering from a disease based on the measurement data acquired in the measurement step 12.

[0076] Therefore, in the disease determination step 16, it is possible to prevent the execution of determinations that result in low reliability compared to when the presence or absence of a disease is always determined regardless of the detection result in the detection step 14.

[0077] Specifically, in the disease determination step 16, the presence or absence of disease is determined based on the measurement data used to calculate the processed value in the calculation step 13, so that it is possible to prevent the execution of determinations that result in unreliable results compared to when always determining the presence or absence of disease, regardless of whether the measurement data is used to calculate the processed value in the calculation step 13.

[0078] <Modification of Calculation Step 13> In the calculation step 13 described above, for each of a plurality of samples, the expression level distribution of the expression levels of a plurality of nucleic acid molecules in the measurement data is compared with a reference distribution to determine a processing value, and a representative processing value is calculated from the determined processing values. However, this is not limited to this. For example, in the calculation step 13, the processing value may be calculated from the sum of the expression levels of all samples.

[0079] Furthermore, in the calculation step 13 described above, the processed value is calculated by a unique digitization process using the measurement data for all expression levels of all samples measured by the same measurement process in the measurement step 12, but this is not limited to this. For example, the processed value may be calculated using measurement data for some expression levels in all samples from the measurement data acquired in the measurement step 12. Furthermore, for example, the processed value may be calculated using measurement data for all expression levels in some samples from the measurement data acquired in the measurement step 12. Furthermore, for example, the processed value may be calculated using measurement data for some expression levels in some samples from the measurement data acquired in the measurement step 12.

[0080] <Modification of Detection Step 14> In the detection step 14 described above, a past processed value calculated before the processed value calculated in the calculation step 13 is used as the reference value, but this is not limited to this. A preset threshold value may also be used as the reference value. Note that when a preset threshold value is used as the reference value, data on the threshold value is recorded in, for example, the storage 34.

[0081] In the above-described detection step 14, a value calculated from processed values ​​in a plurality of consecutive measurements immediately before the measurement in which the processed value was measured is used as the past processed value, but this is not limiting. For example, a value calculated from processed values ​​in a single past measurement may be used as the past processed value.

[0082] Next, examples will be described. Note that the examples are merely examples of the technology of the present disclosure, and the technology of the present disclosure is not limited to the contents of the examples.

[0083] Example 1-1 Calculation of Processing Values ​​In this example, the expression levels of multiple microRNAs for multiple samples were measured by NGS on different measurement days (specifically, the first to seventh measurements (see FIG. 5 )), and seven sets of measurement data indicating the measurement results of the expression levels were obtained. Here, measurement data refers to a group of data indicating the measurement results of the expression levels of multiple microRNAs for multiple samples measured in each of the first to seventh measurements.

[0084] Of this measurement data, the measurement data from the first to fourth measurements were considered to be reliable measurement data with no abnormalities, and the measurement data from the fifth to seventh measurements were verified as measurement data to be detected. Although the cause of the measurement data from the seventh measurement is unknown, measurement data that had been previously confirmed to be abnormal was used as measurement data showing the expression level measurement results by NGS.

[0085] Furthermore, among the measurement data from the first measurement, the measurement data for a specific sample was designated as control data (reference distribution).

[0086] For each of the measurement data from the first to fourth measurements, a treatment value (PSI value) was calculated for each of all samples based on the control data. For each of the first to fourth measurements, the average treatment value of all samples was used as the representative treatment value for that measurement.

[0087] Then, the quartiles of the representative value group from the first to fourth measurements were calculated, and the value of "third quartile + interquartile range x 1.5" (0.0006345391) was used as the reference value for detecting abnormalities.

[0088] For each of the measurement data from the fifth to seventh measurements, a treatment value (PSI value) was calculated for each of all samples based on the control data. For each of the fifth to seventh measurements, the average treatment value of all samples was used as the representative treatment value for that measurement.

[0089] In Fig. 5, the geometric shape shown for each measurement is called a violin plot, which shows the density of data by bulging in the left and right directions. In Fig. 5, the plot shown for each measurement indicates the aforementioned representative value. The dashed line in Fig. 5 indicates the reference value (0.0006345391) for detecting anomalies.

[0090] <Detection of Abnormality> The representative value (0.0002544903) in the fifth measurement did not exceed the reference value (0.0006345391), and therefore it was determined that no abnormality was present.

[0091] The representative value (0.0002884859) in the sixth measurement did not exceed the reference value (0.0006345391), and therefore it was determined that there was no abnormality.

[0092] The representative value (0.0017312941) in the seventh measurement exceeded the reference value (0.0006345391), and was therefore determined to be abnormal, and the determination information was output.

[0093] Example 1-2 Calculation of Treatment Value In this example, measurement data showing the same measurement results as in Example 1-1 was used, and the JSD value was calculated as the treatment value by a method using JSD instead of the method using PSI (see FIG. 6). In this example, the representative value and the reference value were determined under the same conditions as in Example 1-1, except that the treatment value was the JSD value.

[0094] <Detection of abnormality> The representative values ​​in the fifth and sixth measurements did not exceed the reference value, so it was determined that there was no abnormality. However, the representative value in the seventh measurement exceeded the reference value, so it was determined that there was an abnormality, and the determination information was output (see FIG. 6).

[0095] Example 1-3 Calculation of Treatment Value In this example, measurement data showing the same measurement results as in Example 1-1 was used, and the KLD value was calculated as the treatment value by a method using KLD instead of the method using PSI (see FIG. 7). In this example, the representative value and the reference value were calculated under the same conditions as in Example 1-1, except that the treatment value was the KLD value.

[0096] <Detection of abnormality> The representative values ​​in the fifth and sixth measurements did not exceed the reference value, so it was determined that there was no abnormality. However, the representative value in the seventh measurement exceeded the reference value, so it was determined that there was an abnormality, and the determination information was output (see Figure 7).

[0097] Example 1-4 Calculation of Treatment Value In this example, measurement data showing the same measurement results as in Example 1-1 was used, and the EMD value was calculated as the treatment value by a method using EMD instead of the method using PSI (see FIG. 8). In this example, the representative value and the reference value were determined under the same conditions as in Example 1-1, except that the treatment value was the EMD value.

[0098] <Detection of abnormality> The representative values ​​in the fifth and sixth measurements did not exceed the reference value, so it was determined that there was no abnormality. However, the representative value in the seventh measurement exceeded the reference value, so it was determined that there was an abnormality, and the determination information was output (see Figure 8).

[0099] <Example 2> <Calculation of processing value> In this example, the expression levels of multiple microRNAs were measured by NGS on different measurement days (specifically, the first to seventh measurements (see Figure 9)), and seven measurement data showing the measurement results of the expression levels were obtained.

[0100] Of this measurement data, the measurement data from the first to fourth measurements was considered to be reliable measurement data with no abnormalities, and the measurement data from the fifth to seventh measurements was verified as measurement data to be detected. For the measurement data from the seventh measurement, measurement data that had been previously confirmed to be abnormal as measurement data due to a version change in the NGS analysis pipeline was used. In this example, the processing value (PSI value) was calculated under the same conditions as in Example 1-1, and the representative value and reference value were obtained.

[0101] <Detection of Abnormality> The representative values ​​from the fifth and sixth measurements did not exceed the reference value, so it was determined that there was no abnormality, but the representative value from the seventh measurement exceeded the reference value, so it was determined to be abnormal, and the determination information was output (see Figure 9). Note that this abnormality could not be detected using a method using principal component analysis (PCA).

[0102] As described above, it has been demonstrated that anomalies occurring in the process leading up to the acquisition of measurement data can be detected by calculating processed values ​​and detecting anomalies. Similarly, anomalies can also be detected using similar statistical methods other than PSI, EMD, KLD, and JSD. Furthermore, in addition to pipeline version changes in NGS analysis, it has also been confirmed that anomalies caused by fluctuations in the distribution of measurement data that exceed the acceptable range due to reagents used to adjust microRNAs, such as those from different reagent manufacturers, can be detected.

[0103] The present invention is not limited to the above-described embodiment, and various modifications, changes, and improvements are possible without departing from the spirit and scope of the present invention. The above-described modifications may be appropriately combined to form a configuration.

[0104] <Additional Notes> (Aspect 1) A detection method comprising: a measurement step of measuring expression levels of multiple nucleic acid molecules contained in each of multiple samples using the same measurement process and acquiring measurement data indicating the measurement results of the expression levels; a calculation step of calculating a processed value using measurement data for at least some of the multiple samples among the measurement data acquired in the measurement step, by a unique digitization process that converts the multiple expression levels in the measurement data into a unique value; and a detection step of comparing the processed value calculated in the calculation step with a reference value to detect an abnormality that occurred in the process leading up to acquiring the measurement data. (Aspect 2) The detection method according to Aspect 1, wherein the detection step compares the processed value calculated in the calculation step with a previous processed value calculated before the current processed value as the reference value to detect the abnormality. (Aspect 3) The detection method according to Aspect 2, wherein the previous processed value is a value calculated from processed values ​​in multiple measurements earlier than the measurement at which the processed value was measured. (Aspect 4) The detection method according to Aspect 1, wherein the reference value is a preset threshold. (Aspect 5) The detection method according to any one of Aspects 1 to 4, wherein the calculation step comprises: determining a processed value for each of the plurality of samples by comparing the distribution of expression levels of the plurality of nucleic acid molecules in the measurement data with a reference distribution, and calculating a representative value of the processed values ​​from the determined plurality of processed values; and the detection step comprises: comparing the representative value calculated in the calculation step with a reference value to detect an abnormality that occurred in the process leading up to obtaining the measurement data. (Aspect 6) The detection method according to any one of Aspects 1 to 5, wherein the calculation step comprises: calculating the processed value by the unique digitization process using measurement data for all expression levels of all samples measured by the same measurement process in the measurement step. (Aspect 7) The detection method according to any one of Aspects 1 to 6, further comprising: a disease determination step of determining whether or not the donor of the sample has a disease based on the measurement data obtained in the measurement step if no abnormality is detected in the detection step.(Aspect 8) The detection method according to Aspect 7, wherein the disease assessment step determines the presence or absence of the disease based on the measurement data used to calculate the processed value in the calculation step. (Aspect 9) The detection method according to any one of Aspects 1 to 8, wherein the measurement step measures expression levels of microRNAs as multiple nucleic acid molecules contained in each of the multiple samples by the same measurement process using a next-generation sequencer. (Aspect 10) A detection device comprising a processor, wherein the processor calculates a processed value by using measurement data for at least some of the multiple samples among measurement data indicating measurement results of the expression levels of multiple nucleic acid molecules contained in each of the multiple samples, the measurement data being measured in advance by the same measurement process, by a unique digitization process that converts the multiple expression levels in the measurement data into a unique value, and compares the calculated processed value with a reference value to detect an abnormality that occurred in the process of obtaining the measurement data. (Aspect 11) A detection system comprising: a measurement unit that measures expression levels of multiple nucleic acid molecules contained in each of multiple samples by the same measurement process and acquires measurement data indicating the measurement results of the expression levels, a calculation unit that calculates a processed value using measurement data for at least some of the multiple samples from the measurement data acquired by the measurement unit by unique digitization processing that converts the multiple expression levels in the measurement data into a unique value, and a detection unit that compares the processed value calculated by the calculation unit with a reference value to detect an abnormality that occurred in the process up to acquiring the measurement data. (Aspect 12) A detection program that causes a computer to execute a detection process that calculates a processed value using measurement data for at least some of the multiple samples from measurement data indicating measurement results of the expression levels, the expression levels of multiple nucleic acid molecules contained in each of multiple samples, measured in advance by the same measurement process, by unique digitization processing that converts the multiple expression levels in the measurement data into a unique value, and compares the calculated processed value with a reference value to detect an abnormality that occurred in the process up to acquiring the measurement data.(Aspect 13) A non-transitory recording medium having recorded thereon a detection program for causing a computer to execute a detection process, which comprises: calculating a processed value by a unique digitization process that converts the multiple expression levels in the measurement data into a unique value using measurement data for at least some of multiple samples, the measurement data indicating the expression levels of multiple nucleic acid molecules contained in each of multiple samples, the measurement data having been previously measured by the same measurement process; and comparing the calculated processed value with a reference value to detect abnormalities that occurred in the process leading up to obtaining the measurement data.

[0105] The disclosure of Japanese Patent Application No. 2023-187040, filed on October 31, 2023, is incorporated herein by reference in its entirety. All documents, patent applications, and technical standards mentioned herein are incorporated herein by reference to the same extent as if each individual document, patent application, and technical standard was specifically and individually indicated to be incorporated by reference.

Claims

1. A detection method comprising: a measurement step of measuring the expression levels of multiple nucleic acid molecules contained in each of multiple samples by the same measurement process, and obtaining measurement data indicating the measurement results of the expression levels; a calculation step of calculating a processed value using the measurement data for at least a portion of the multiple samples obtained in the measurement step by a unique numerical conversion process that converts the multiple expression levels in the measurement data into a unique value; and a detection step of comparing the processed value calculated in the calculation step with a reference value to detect an abnormality that occurred in the process leading up to obtaining the measurement data.

2. The detection method according to claim 1, wherein the detection step detects the abnormality by comparing the processing value calculated in the calculation step with a past processing value calculated prior to the processing value in question as the reference value.

3. The detection method according to claim 2, wherein the past processed value is a value calculated from processed values ​​in a plurality of measurements prior to the measurement at which the processed value was measured.

4. The detection method according to claim 1, wherein the reference value is a preset threshold value.

5. The detection method according to claim 1, wherein in the calculation step, a processing value is obtained for each of the plurality of samples by comparing the distribution of expression levels of the plurality of nucleic acid molecules in the measurement data with a standard distribution, and a representative value of the processing values ​​is calculated from the obtained plurality of processing values; and in the detection step, the representative value calculated in the calculation step is compared with a standard value to detect an abnormality that occurred in the process leading up to obtaining the measurement data.

6. The detection method according to claim 1, wherein the calculation step calculates the processed value by the unique quantification process using measurement data for all expression levels of all samples measured by the same measurement process in the measurement step.

7. The detection method according to claim 1, further comprising a disease determination step of determining whether or not the donor of the sample has a disease based on the measurement data obtained in the measurement step if the abnormality is not detected in the detection step.

8. The detection method according to claim 7, wherein the disease determination step determines the presence or absence of the disease based on the measurement data used to calculate the processed value in the calculation step.

9. The detection method according to claim 8, wherein in the measurement step, the expression levels of microRNA as multiple nucleic acid molecules contained in each of the multiple samples are measured by the same measurement process using a next-generation sequencer.

10. A detection device comprising a processor, which calculates a processed value using measurement data for at least a portion of a plurality of samples that indicates the expression levels of a plurality of nucleic acid molecules contained in each of a plurality of samples and that are measurement results of the expression levels previously measured by the same measurement process, by a unique numerical conversion process that converts the plurality of expression levels in the measurement data into a unique value, and compares the calculated processed value with a reference value to detect an abnormality that occurred in the process of obtaining the measurement data.

11. A detection system comprising: a measurement unit that measures the expression levels of multiple nucleic acid molecules contained in each of multiple samples using the same measurement process and acquires measurement data indicating the measurement results of the expression levels; a calculation unit that calculates a processed value using the measurement data for at least a portion of the multiple samples acquired by the measurement unit by a unique numerical conversion process that converts the multiple expression levels in the measurement data into a unique value; and a detection unit that compares the processed value calculated by the calculation unit with a reference value to detect abnormalities that occurred in the process leading up to acquiring the measurement data.

12. A detection program for causing a computer to execute a detection process which calculates a processed value by a unique quantification process which converts the multiple expression amounts in the measurement data into a unique value using measurement data for at least a portion of a plurality of samples, the measurement data indicating the expression amounts of a plurality of nucleic acid molecules contained in each of a plurality of samples and which have been previously measured by the same measurement process, and compares the calculated processed value with a reference value to detect an abnormality which occurred in the process leading up to obtaining the measurement data.

13. A non-transient recording medium having recorded thereon a detection program for causing a computer to execute a detection process which uses measurement data for at least a portion of a plurality of samples, the measurement data indicating the expression levels of a plurality of nucleic acid molecules contained in each of a plurality of samples and the measurement results of the expression levels previously measured by the same measurement process, to calculate a processed value by a unique quantification process which converts the plurality of expression levels in the measurement data into a unique value, and compares the calculated processed value with a reference value to detect an abnormality which occurred in the process leading up to obtaining the measurement data.

Images