Method of preparing and method of storing blood sample for small RNA measurement, method of measuring small RNA expression level, disease assessment method, evaluation method, model generation method, evaluation device, evaluation system, evaluation program, and recording medium

Abstract

Provided is a method of preparing a blood sample for small RNA measurement, said method involving separating blood serum or blood plasma from blood collected from a living body and adding a preservative to the blood serum or blood plasma within four hours of the collection of the blood. Also provided are practical applications for the method.

Description

Methods for preparing and storing blood samples for measuring small RNAs, methods for measuring expression levels of small RNAs, methods for determining disease, testing methods, methods for generating models, testing devices, testing systems, testing programs, and recording media

[0001] The present disclosure relates to methods for preparing and storing blood samples for measuring small RNAs, methods for measuring the expression levels of small RNAs, methods for diagnosing diseases, testing methods, methods for generating models, testing devices, testing systems, testing programs, and recording media.

[0002] Small RNAs are small molecules of RNA present in living organisms, and their structure and function are being analyzed. Small RNAs are known to be associated with various diseases, including cancer and cardiovascular disease, and investigating the dynamics of small RNAs in living organisms may be useful for the diagnosis and treatment of diseases.

[0003] Various techniques are being developed for preparing nucleic acid samples, measuring the expression levels of nucleic acids, and diagnosing diseases based on the expression levels of small RNAs.

[0004] Patent Document 1 describes a disease prevalence determination device that includes a sample data acquisition unit that acquires sample data including the expression levels of multiple types of microRNA in a biological sample, and a prevalence determination unit that uses a trained model that can determine the prevalence of multiple diseases to output prevalence determination results for multiple diseases in multiple body parts for the acquired sample data.

[0005] Patent Document 2 describes a method for purifying nucleic acids present in an in vivo cellular environment, which comprises mixing a medium containing nucleic acids with a binding matrix and a compound, the compound comprising guanidine thiocyanate and acetamide and / or an acetamide derivative.

[0006] Patent Document 3 describes a method for preserving nucleic acids in a body fluid, which comprises adding to the body fluid a nucleic acid preservation solution containing a divalent metal chelator such as ethylenediaminetetraacetic acid (EDTA) and a chelator-enhancing component such as lithium chloride or guanidine.

[0007] Patent Document 4 describes a method for preventing the degradation of RNA derived from lipids on the skin surface, which comprises treating a collection medium containing lipids on the skin surface with an aqueous solution containing at least one selected from the group consisting of guanidine hydrochloride and guanidine thiocyanate, with water or an aqueous alcohol solution as the solvent.

[0008] Patent Document 5 describes a stabilizing buffer containing guanidine hydrochloride, Triton X-100, and EDTA and having a pH of less than 4.5. It also describes measuring RNA expression levels using a biological sample stabilized with the stabilizing buffer, and diagnosing or treating diseases based on the RNA expression levels.

[0009] Patent Document 6 describes a nucleic acid preservation composition comprising a carrier, a chaotropic agent, a buffer, a chelating agent, a detergent or surfactant, an alcohol, and a mucolytic or reducing agent.

[0010] Patent Document 7 describes a container for collecting and storing blood, which contains an aqueous solution having a pH of 4 to 7.5 and which contains a guanidinium salt, a buffer, a reducing agent, and a surfactant.

[0011] Non-Patent Document 1 describes a method for processing microRNAs in which, after blood collection, the blood is coagulated, centrifuged, and serum is collected in a separate container, and the serum is stored at -80°C.

[0012] Japanese Patent No. 7021097 Japanese Patent No. 5777221 Special Publication No. 2001-526051 International Publication No. 2022 / 045303 Special Publication No. 2022-553914 Japanese Patent Laid-Open No. 2022-69447 Japanese Patent No. 4607323

[0013] Louise C. Laurent et al., "Serum Collection Procedure (Small Scale) for the analysis of extracellular RNA", Research Square (USA), December 7, 2015, pp. 1-3

[0014] In preparing blood samples for measuring small RNAs, the measured values ​​of small RNA expression levels vary depending on the preparation and storage conditions of the blood samples at the blood collection site, so the preparation and storage methods of the blood samples may affect the measurement results of small RNA expression levels and, ultimately, the disease assessment results. In light of the above circumstances, the present disclosure relates to methods for preparing and storing blood samples for measuring small RNAs, methods for measuring small RNA expression levels, testing methods, model generation methods, testing devices, testing systems, testing programs, and recording media, which enable high-precision measurement of small RNA expression levels, as well as disease assessment methods using these.

[0015] Means for solving the above problems include the following aspects. <1> A method for preparing a blood sample for measuring small RNAs, comprising: separating serum or plasma from blood collected from a living body; and adding a preservative to the serum or plasma within four hours of collecting the blood. <2> The method according to <1>, wherein the time between collecting the blood and adding the preservative is within two hours. <3> The method according to <1> or <2>, wherein the time between collecting the blood and adding the preservative is within one hour. <4> The method according to any one of <1> to <3>, wherein the preservative comprises a guanidinium salt. <5> The method according to any one of <1> to <4>, wherein separating the serum or plasma includes centrifuging the blood, and wherein the time between collecting the blood and completing the centrifugation is within two hours. <6> The method according to <5>, wherein the time between collecting the blood and completing the centrifugation is within one hour. <7> A method for storing a blood sample for measuring small RNAs, comprising storing a blood sample prepared by the method according to any one of <1> to <6>. <8> A method for measuring the expression levels of small RNAs, comprising: preparing a blood sample in which a preservative has been added to serum or plasma separated from blood within 4 hours of collecting the blood; and measuring the expression levels of multiple small RNAs in the blood sample. <9> A method for determining the presence or absence of a disease in a subject using small RNA data indicating the expression levels of small RNAs obtained by the method described in <8>, comprising determining the presence or absence of the disease using a discrimination criterion based on the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease incidence data indicating the presence or absence of the disease in the multiple subjects.<9-2> A method for testing expression levels of small RNAs, comprising measuring the expression levels of multiple small RNAs in a test subject by the method described in <8>, wherein the expression levels of the small RNAs indicate the presence or absence of a disease in the test subject based on a discrimination criterion based on the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and incidence data showing the presence or absence of a disease in the multiple subjects. <10> The method described in <9>, wherein the presence or absence of a disease is determined by inputting the small RNA data of the test subject into a disease-trained model that has been machine-learned to determine the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and incidence data showing the presence or absence of a disease in the multiple subjects. <11> A testing method for testing a preservative addition time from blood collection to addition of a preservative, using small RNA data showing the results of measuring the expression levels of multiple small RNAs in a blood sample, the testing method comprising testing the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition times for the multiple blood samples. <12> A method for generating a preservative addition time trained model used in a testing method for testing the preservative addition time from blood collection to the addition of a preservative, the testing method performing the preservative addition time by inputting small RNA data showing the results of measuring the expression levels of multiple small RNAs in blood samples into a preservative addition time trained model, the method comprising: generating the preservative addition time trained model by machine learning the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition times for the multiple blood samples.<13> A method for diagnosing a disease in a subject using small RNA data in a blood sample for which the test method according to <11> has given a test result indicating that the preservative addition time is appropriate, the method comprising determining the presence or absence of the disease by inputting the small RNA data of the subject into a disease-trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data indicating the presence or absence of the disease in the multiple subjects. <14> A method for generating a disease determination model for diagnosing a disease in a subject based on the expression levels of multiple small RNAs, the method comprising selecting a specimen from which multiple small RNAs used to generate the disease determination model are derived, the method comprising selecting a specimen for which the preservative addition time between collection of the specimen from the subject and addition of the preservative is appropriate as a standard specimen, and generating the disease determination model using the expression levels of the multiple small RNAs in the standard specimen. <15> The method according to <14>, wherein the standard specimen is a specimen corresponding to small RNA data in a blood sample for which a test result indicating that the preservative addition time is appropriate has been obtained by the test method according to <11>. <16> A testing device that tests the preservative addition time from blood collection to the addition of a preservative using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in a blood sample, the testing device comprising a processor, wherein the processor tests the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition times for the multiple blood samples.<17> A testing system that tests the preservative addition time from blood collection to the addition of a preservative, using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in a blood sample, comprising: a measurement unit that obtains the small RNA data; and a testing unit that tests the preservative addition time by inputting the small RNA data obtained by the measurement unit into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time for the multiple blood samples. <18> A testing program for causing a computer to execute a testing process for testing a preservative addition time from blood collection to addition of a preservative, using small RNA data showing the results of measuring the expression levels of multiple small RNAs in blood samples, by inputting the small RNA data into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition times for the multiple blood samples. <19> A non-transitory recording medium having recorded thereon a testing program for causing a computer to execute a testing process for testing a preservative addition time from blood collection to addition of a preservative, using small RNA data showing the results of measuring the expression levels of multiple small RNAs in blood samples, by inputting the small RNA data into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition times for the multiple blood samples.

[0016] According to the present disclosure, there are provided a method for preparing and storing blood samples for measuring small RNAs, a method for measuring the expression level of small RNAs, a testing method, a model generation method, a testing device, a testing system, a testing program, and a recording medium, which enable highly accurate measurement of the expression level of small RNAs, as well as a method for diagnosing diseases using these.

[0017] Fig. 1 is a schematic block diagram of an example of a computer that functions as an examination apparatus according to an embodiment of the present disclosure. Fig. 2 is a block diagram showing the functional configuration of the examination apparatus according to an embodiment of the present disclosure. Fig. 3 shows prediction scores for blood samples stored 2 hours, 4 hours, 6 hours, and 24 hours after the start of the test, a negative control (without the addition of a preservative solution), and a positive control (stored at -80°C), using the lung cancer discrimination model created in the examples.

[0018] Hereinafter, modes for carrying out embodiments of the present disclosure will be described in detail. However, the embodiments of the present disclosure are not limited to the following embodiments. In the following embodiments, components (including element steps, etc.) are not essential unless otherwise specified. The same applies to numerical values ​​and their ranges, and they do not limit the embodiments of the present disclosure.

[0019] In the present disclosure, the term "process" includes not only processes that are independent of other processes, but also processes that cannot be clearly distinguished from other processes as long as the purpose of the process is achieved. In the present disclosure, numerical ranges indicated using "to" include the numerical values ​​before and after "to" as the minimum and maximum values, respectively. In numerical ranges described in stages in the present disclosure, the upper or lower limit value described in one numerical range may be replaced with the upper or lower limit value of another numerical range described in stages. Furthermore, in numerical ranges described in the present disclosure, the upper or lower limit value of that numerical range may be replaced with a value shown in the examples. When describing embodiments of the present disclosure with reference to the drawings, 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, descriptions of configurations not directly related to the present disclosure or well-known configurations may be omitted.

[0020] <Method for Preparing Blood Samples for Small RNA Measurement> In one embodiment of the present disclosure, a method for preparing blood samples for small RNA measurement includes separating serum or plasma from blood collected from a living body and adding a preservative to the serum or plasma within 4 hours of collecting the blood. In one aspect, the preparation method of this embodiment is carried out based on instructions for carrying out the method.

[0021] The inventors have found that adding a preservative to a plasma sample for measuring small RNA expression levels within 4 hours of blood collection reduces the variability in measurement values ​​associated with sample preparation and enables highly accurate measurement of small RNA expression levels. That is, a correlation was found between the time from blood collection to preservative addition (also referred to as "preservative addition time" in the present disclosure) and the variability in the measurement values ​​of small RNA expression levels, and it was found that controlling the time until preservative addition within a certain period enables highly accurate measurement.

[0022] If the preservative addition time is prolonged, the blood sample may deteriorate, potentially changing the measurement results in the small RNA measurement process. For example, if the preservative addition time is prolonged, small RNAs such as microRNAs contained in the blood sample may be more easily degraded, which may result in a decrease in the expression level of the small RNA measured in the measurement process. Furthermore, if the preservative addition time is prolonged, the small RNA is degraded, particularly starting from the terminal base. For example, if the 3' base of miR-16-5p (UAGCAGCACGUAAAUAUUGGCG (SEQ ID NO: 1)) is degraded, a two-base mismatch occurs with miR-195-5p (UAGCAGCACAGAAAUAUUGGC (SEQ ID NO: 2)). If the base sequence identification in the measurement process allows for a two-base mismatch, miR-16-5p may be recognized as miR-195-5p. In this way, degradation of small RNAs may change the amount of small RNA present in the measurement process. This phenomenon becomes more pronounced when the preservative addition time exceeds a predetermined time, affecting the measurement results of small RNA measured in the measurement step. According to the preparation method of this embodiment, by adding the preservative within a specific time, it is thought that the above-mentioned erroneous recognition can be reduced and high-precision measurement can be achieved.

[0023] The blood sample obtained by the preparation method of this embodiment also tends to be highly resistant to freezing and thawing and to temperature changes associated with transportation, etc. In other words, the preparation method of this embodiment widens the allowable range of storage temperatures.

[0024] Examples of small RNAs to be measured include microRNA, piRNA, tsRNA, etc. The number of bases in the small RNA is not particularly limited, and it is preferably 1 to 100 bases long, more preferably 10 to 50 bases long, and even more preferably 16 to 40 bases long. Small RNAs are typically measured using multiple types of small RNAs in a blood sample.

[0025] [Separation Step] The method for preparing a blood sample for measuring small RNAs includes separating serum or plasma from blood collected from a living body (also referred to as the "separation step" in the present disclosure). The subject from which blood is collected may be a human or a non-human animal. 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.

[0026] Separation of serum or plasma can be carried out using known methods. Typically, separation of serum or plasma involves centrifuging blood. For example, serum separation can be carried out by collecting blood using a blood collection tube containing a serum separating agent, coagulating the blood at room temperature (e.g., 15°C to 40°C, specifically about 25°C), and then centrifuging. Furthermore, plasma separation can be carried out by collecting blood using a blood collection tube containing an anticoagulant (e.g., heparin; ethylenediaminetetraacetic acid (EDTA), calcium chelating agents such as sodium citrate, etc.), and then centrifuging. When separating serum, the time for coagulating the blood by leaving it at room temperature (e.g., 15°C to 40°C, specifically about 25°C) is preferably 15 to 120 minutes, more preferably 20 to 90 minutes, and even more preferably 30 to 60 minutes.

[0027] In one embodiment, when collecting serum, the collection process is carried out, for example, as follows. First, blood is collected using a vacuum blood collection tube containing a serum separating agent (blood collection process). Next, immediately after the blood collection process, the blood is mixed by inversion. Next, the blood is left at room temperature (e.g., 15°C to 40°C, specifically about 25°C) for 30 minutes or more to allow it to clot. Next, the blood is centrifuged and the serum is separated (centrifugation process).

[0028] In separating serum or plasma, the time from blood collection to the completion of centrifugation (specifically, the time from blood collection to separation of serum or plasma into another container after centrifugation) is preferably within 2 hours, more preferably within 1 hour, and even more preferably within 30 minutes, from the viewpoint of shortening the time until addition of a preservative.

[0029] [Preservative Addition Step] The method for preparing a blood sample for measuring small RNAs includes adding a preservative to the separated serum or plasma within 4 hours of blood collection (in the present disclosure, the step of adding a preservative to the separated serum or plasma is also referred to as the "preservative addition step"). The preservative is added to the separated serum or plasma within 4 hours of blood collection. From the viewpoint of measuring the expression level of small RNAs with higher accuracy, the preservative addition time is preferably within 2 hours, and more preferably within 1 hour.

[0030] Preservatives are typically liquid and contain preservative components for stably preserving small RNAs. Preservatives may contain other additives in addition to the preservative components. Examples of preservative components include chaotropic agents, reducing agents, buffers, surfactants, and chelating agents. Examples of chaotropic agents include guanidinium salts, urea, and iodide salts. Examples of guanidinium salts include guanidinium thiocyanate, guanidinium thioisocyanate, guanidine hydrochloride, guanidinium chloride, guanidinium iodide, sodium iodide, and potassium iodide. Examples of reducing agents include thioglycerin, mercaptoethanol, dithiothreitol, tri(2-carboxyethyl)phosphine, and dithiothreitol (DTT). Examples of buffers include sodium acetate, sodium citrate, sodium bicarbonate, tris(hydroxymethyl)aminomethane, sodium hydrogen phosphate, sodium dihydrogen phosphate, and so-called Good's buffers. Surfactants include sodium lauroyl sarcosinate, sodium dodecyl sulfate, polysorbate (Tween™), lauryl dimethylamine oxide, cetyltrimethylammonium bromide, polyethoxylated alcohol, polyoxyethylene, sorbitan, octoxynol (Triton X-100™), N,N-dimethyldodecylamine-N-oxide, hexadecyltrimethylammonium bromide, polyoxyl 10 lauryl ether, sodium deoxycholate, sodium cholate, polyoxyl castor oil (Cremophor™), nonylphenol ethoxylate (Tergitol™), cyclodextrin, lecithin, and methylbenzethonium chloride (Hyamine™). Chelating agents include EDTA, citric acid, phytic acid, and gluconic acid. Other additives include methylacetamide. Each component contained in the preservative may be used alone or in combination of two or more.

[0031] In particular, the preservative preferably contains a guanidinium salt. Guanidinium salts have the effect of inactivating RNases present in biological samples such as serum, thereby effectively stabilizing RNA. Guanidinium salts are also useful in that they can detoxify the infectivity of serum and the like.

[0032] In one embodiment, the preservative may include a chaotropic agent, a reducing agent, a buffer, and N-methylacetamide, and may not include other components. In a further embodiment, the preservative may include a chaotropic agent, a reducing agent, and a buffer, and may not include other components. In one embodiment, the preservative may not include a chelating agent (e.g., EDTA). The preparation method of the present disclosure enables high-accuracy measurement of small RNAs using a relatively simple formulation.

[0033] <Method for storing blood samples for small RNA measurement> In one embodiment of the present disclosure, a method for storing blood samples for small RNA measurement includes storing a blood sample prepared by the above-described preparation method. The storage method according to this embodiment enables highly accurate measurement of the expression level of small RNA. The blood sample may be stored at a predetermined storage temperature while being transported, for example, to a measurement location where a subsequent measurement will be performed. The storage method according to this embodiment enables highly accurate measurement of the expression level of small RNA, even if the blood sample is stored for a certain period of time for transportation from the blood collection site to the measurement location.

[0034] Storage conditions for blood samples can be selected appropriately. For example, the storage temperature is preferably 30°C or lower, more preferably 10°C or lower, even more preferably -4°C or lower, and particularly preferably -50°C or lower. For example, blood samples may be stored at 4°C, -20°C, or -80°C. The storage temperature may be -100°C or higher, or may be -50°C or higher. Therefore, the storage temperature is preferably -100°C to 30°C.

[0035] The storage period of the blood sample is not particularly limited and can be appropriately selected depending on the storage conditions, etc. For example, the storage period is preferably within 30 days, more preferably within 20 days, even more preferably within 10 days, and particularly preferably within 5 days. Meanwhile, according to the above-described blood sample preparation method, the expression level of small RNA can be measured with high accuracy by adding a preservative within 4 hours of blood collection. From this perspective, the storage period may be 1 day or more, 3 days or more, or 5 days or more. The blood sample may be stored at a predetermined storage temperature and transported, for example, to a measurement location where a subsequent measurement will be performed. Therefore, the storage period is preferably 0 to 30 days, and may be 1 to 30 days.

[0036] <Method for measuring expression levels of small RNAs> A method for measuring expression levels of small RNAs in one embodiment of the present disclosure includes preparing a blood sample in which a preservative is added to serum or plasma separated from blood within 4 hours of collecting the blood, and measuring the expression levels of multiple small RNAs in the blood sample. Hereinafter, the step of measuring the expression levels of multiple small RNAs in a blood sample will also be referred to as the "measurement step." The method for measuring expression levels of small RNAs according to this embodiment enables highly accurate measurement of the expression levels of small RNAs. Details of the small RNAs to be measured and the preparation of the blood sample to which a preservative has been added are as described above.

[0037] Methods for measuring the expression level of small RNAs include next-generation sequencers (NGS), DNA chips, quantitative PCR, flow cytometry, etc. Before measuring the expression level of small RNAs, pretreatment for measurement may be performed as necessary. If the expression levels of multiple small RNAs can be measured and the results (i.e., small RNA data) can be obtained, various methods, including known methods, can be used to measure the expression levels of multiple small RNAs. The small RNA data may be absolute quantitative values ​​quantified as absolute values, or relative expression levels quantified as relative expression levels.

[0038] In one embodiment, the expression level of small RNAs is measured by NGS. NGS allows the expression levels of multiple small RNAs in multiple blood samples to be measured simultaneously. Measurement of small RNA expression levels by NGS is performed, for example, as follows: Multiple small RNAs contained in a blood sample are measured by NGS, and the base sequence of each small RNA is identified. Next, the number of identified small RNAs is counted for each base sequence to determine the number of small RNA reads in NGS. This number of small RNA reads corresponds to the expression level of the small RNA (specifically, the absolute expression level). Note that the expression levels of multiple small RNAs (e.g., microRNAs) in a blood sample are absolute values ​​derived from the living body, but because the expression levels must be quantified through a process involving measurement devices and reagent treatment, it is difficult to quantify the expression levels of small RNAs in a blood sample as absolute values. Therefore, the expression levels of small RNAs may be expressed as relative values ​​by data processing of the NGS measurement results, and even such relative values ​​represent small RNA data.

[0039] In addition, the number of reads of small RNAs output by NGS may be normalized and expressed as relative expression levels (i.e., relative expression levels). Normalization methods that can be used include RPM (Read Per Million) normalization and normalization using an internal standard small RNA. In this way, small RNA data may represent normalized relative expression levels.

[0040] <Disease Determination Method 1> In one embodiment of the present disclosure, the disease determination method is a method for determining the disease of a subject using small RNA data indicating the expression levels of small RNAs obtained by the aforementioned measurement method, and includes determining the presence or absence of the disease using a discrimination criterion based on the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data indicating the presence or absence of the disease in the multiple subjects. Hereinafter, the step of determining the presence or absence of the disease is also referred to as "disease determination step 1." According to the disease determination method of this embodiment, disease determination is performed using small RNAs obtained by the aforementioned measurement method, thereby reducing erroneous determination and enabling highly accurate disease determination.

[0041] The subject may be a human or a non-human animal, including non-human mammals (monkeys, dogs, cats, mice, rats, rabbits, cows, horses, pigs, sheep, etc.), birds (chickens, quails, etc.), etc.

[0042] Diseases to be assessed include cancer, diabetes, and the like. Cancers include solid cancers and blood cancers. Solid cancers include head and neck cancer, thyroid cancer, breast cancer, lung cancer, pancreatic cancer, colon cancer, hepatocellular carcinoma, biliary tract cancer, stomach cancer, esophageal cancer, small intestine cancer, rectal cancer, colon cancer, kidney cancer, bladder cancer, testicular cancer, prostate cancer, ovarian cancer, uterine cancer, cervical cancer, skin cancer, brain tumor, malignant melanoma, osteosarcoma, and the like. Blood cancers include leukemia (e.g., acute myeloid leukemia, chronic myeloid leukemia, acute lymphocytic leukemia, and chronic lymphocytic leukemia), lymphoma (e.g., Hodgkin's lymphoma and non-Hodgkin's lymphoma), multiple myeloma, and the like. The disease to be assessed may be one type or multiple types.

[0043] Examples of discrimination criteria based on the correlation between measurement data and disease data include discrimination criteria based on a disease learning model that uses machine learning to learn the correlation, discriminant equations based on multiple small RNA measurement data, and discriminant equations generated by combining small RNA measurement data with general clinical test data such as various tumor marker data and biochemical test data.

[0044] In one embodiment, the presence or absence of a disease is determined by inputting small RNA data of a test subject into a disease-trained model that has been machine-learned to determine the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data showing the presence or absence of the disease in the multiple subjects. A method for generating a disease-trained model is described in detail below.

[0045] - Method for generating a disease-trained model - The disease-trained model is generated by machine learning the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data showing the presence or absence of a disease in the multiple subjects. The multiple subjects typically include subjects who do not have the disease to be assessed and subjects who have the disease to be assessed.

[0046] The measurement data is data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects. As described above, the measurement data can be measured by various methods, including known methods such as next-generation sequencers (NGS), DNA chips, quantitative PCR, and flow cytometry.

[0047] Measurement data used for machine learning of correlations may be measurement data of small RNAs with high expression levels in multiple blood samples. That is, measurement data excluding small RNAs with relatively low expression levels may be used. For example, measurement data of a predetermined number of small RNAs whose expression levels are ranked in the top 20 to 300, 30 to 200, or 50 to 150, for example, up to the top 100, in multiple blood samples may be used. Small RNAs with relatively low expression levels have a large fold-change even with small fluctuations in expression level, making them more likely to function as explanatory variables. However, they are susceptible to fluctuations due to factors other than the element of interest (specifically, the preservative addition time) and fluctuations due to measurement errors, resulting in a drawback of weak robustness. Therefore, the effect of enhancing robustness can be achieved by using only small RNAs with relatively high expression levels as training data.

[0048] The measurement data used for machine learning of correlation may be measurement data excluding small RNAs whose expression level in any of the blood samples was zero. Since fluctuations from zero result in infinite fold changes and have a large impact, it is preferable to exclude small RNAs whose expression level in any of the blood samples used as training data from the training data.

[0049] The measurement data used for the machine learning of correlations is preferably obtained using blood samples prepared by the above-mentioned blood sample preparation method. That is, the measurement data is preferably obtained using blood samples to which a preservative has been added within 4 hours of blood collection. This makes it possible to build a model that removes noise caused by deterioration of the blood sample.

[0050] The morbidity data is data indicating whether or not the plurality of subjects have a disease, and typically includes whether or not the subjects have the disease to be determined.

[0051] A disease-trained model is created by machine learning the correlation between these measurement data and disease data. For machine learning, various linear and nonlinear algorithms known as machine learning algorithms can be used, or multiple algorithms can be combined. For example, the following algorithms can be used: Random forest, Gradient boosted trees, Extreme gradient boosted trees, Light-gradient boosted machine, Neural networks, Regularized regression, Elastic-net, K-Nearest neighbors, Support vector machine, Generalized additive model

[0052] In one embodiment, the disease trained model may be generated by qualitatively training a learning model to determine whether or not a subject has a disease. For example, the disease trained model is generated using, as training data, first measurement data obtained from a blood sample derived from a subject with the disease and second measurement data obtained from a blood sample derived from a subject without the disease. In this case, the presence or absence of a disease is determined by inputting RNA data derived from the subject into the disease trained model. Therefore, in the disease determination step 1, a determination result indicating that the subject has the disease or that the subject does not have the disease is displayed. If there are multiple types of diseases to be determined, a determination result indicating that the subject has the disease or that the subject does not have the disease is displayed for each of the multiple types of diseases.

[0053] In one embodiment, the disease trained model may be generated by quantitatively training a learning model on a disease. In this case, the disease trained model determines, for example, the probability of disease onset, and outputs the determination result as a judgment value. Then, based on a comparison between the output judgment value and a threshold, the presence or absence of the disease is determined.

[0054] In a further embodiment, the present disclosure also provides a method for testing expression levels of small RNAs, comprising measuring the expression levels of multiple small RNAs in a test subject by the aforementioned measurement method, wherein the expression levels of the small RNAs indicate the presence or absence of a disease in the test subject based on a discrimination criterion based on the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data showing the presence or absence of a disease in the multiple subjects. Details of this embodiment are as described in "Disease Determination Method 1."

[0055] <Method for Testing Preservative Addition Time> A method for testing preservative addition time in one embodiment of the present disclosure is a method for testing the preservative addition time from blood collection to the addition of a preservative using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in blood samples, and includes testing the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has been trained by machine learning to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition times for the multiple blood samples. Hereinafter, the step of performing the test is also referred to as the "testing step."

[0056] According to the testing method of this embodiment, for example, it is possible to determine whether the preservative addition time was appropriate based solely on data related to the expression level of small RNA in the blood sample to be measured. Therefore, for example, there is no need to record data such as work records at the blood collection site, and the quality of the blood sample can be determined independently of the blood collection site during measurement. Furthermore, according to the testing method of this embodiment, erroneous diagnosis due to poor sample quality can be prevented in advance when diagnosing a disease. On the other hand, since a tolerance range can be set that does not significantly affect disease diagnosis, there is also the advantage of reducing the burden on the blood collection site.

[0057] In the test of the preservative addition time, whether the preservative addition time is appropriate or not is examined. For example, it is determined whether the preservative addition time is within a predetermined time (within 4 hours, preferably within 2 hours, more preferably within 1 hour), and if the preservative addition time is within the predetermined time, it is determined as appropriate (good), and if it exceeds the predetermined time, it is determined as inappropriate (bad).

[0058] The blood sample may be serum or plasma. Details of the method for separating serum or plasma are as described above. Details of the preservative are as described above. Details of the small RNA data are as described above.

[0059] The method for selecting subjects from which multiple blood samples for obtaining measurement data are derived is not particularly limited, and may be randomly selected. For example, subjects from which the multiple blood samples are derived may be randomly selected regardless of the presence and / or type of disease. The multiple blood samples may include blood samples from healthy individuals or blood samples from diseased individuals. The multiple blood samples may consist solely of blood samples from healthy individuals or blood samples from diseased individuals. When disease is determined in a subsequent step, it is preferable that some or all of the blood samples are derived from patients with the same disease as the disease to be determined. Examples of diseases include those described in the "Disease Determination Method 1" section. In one embodiment, the multiple blood samples may include blood samples from cancer patients or may consist solely of blood samples from cancer patients. Examples of cancer include those described in the "Disease Determination Method 1" section.

[0060] As described above, the expression levels of small RNAs can be measured by various techniques, including known techniques such as next-generation sequencers (NGS), DNA chips, quantitative PCR, flow cytometry, etc. The method for acquiring small RNA data is as described above.

[0061] In this embodiment, small RNA data for a blood sample to be tested is input into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition times for the multiple blood samples. The preservative addition time trained model can be generated by the following method.

[0062] -Method for generating a preservative addition time trained model- A method for generating a preservative addition time trained model in one embodiment of the present disclosure is a method for generating a preservative addition time trained model used in a testing method for testing the preservative addition time from blood collection to the addition of a preservative, the testing method for testing the preservative addition time by inputting small RNA data indicating the results of measuring the expression levels of multiple small RNAs in blood samples into a preservative addition time trained model, the method comprising: generating the preservative addition time trained model by machine learning the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition times in the multiple blood samples.

[0063] The preservative addition time trained model is generated by machine learning the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition times in the multiple blood samples.

[0064] The measurement data is data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples. The measurement data can be measured using multiple blood samples by various techniques, including known techniques such as next-generation sequencers (NGS), DNA chips, quantitative PCR, and flow cytometry, as described above.

[0065] The measurement data may be of small RNAs with high expression levels in multiple blood samples. That is, measurement data excluding small RNAs with relatively low expression levels may be used. For example, measurement data of a predetermined number of small RNAs whose expression levels are in the top 20 to 300, 30 to 200, or 50 to 150, for example, up to the top 100, may be used in multiple blood samples.

[0066] The measurement data may be obtained by excluding small RNAs whose expression level is zero in any of the blood samples. Since the fold change from zero is infinite and has a large impact, it is preferable to exclude small RNAs whose expression level is zero in any of the samples used as training data from the training data.

[0067] The measurement data used for the machine learning of correlations is preferably obtained using blood samples prepared by the above-mentioned blood sample preparation method. That is, the measurement data is preferably obtained using blood samples to which a preservative has been added within 4 hours of blood collection. This makes it possible to build a model that removes noise caused by deterioration of the blood sample.

[0068] The preservative addition time data is data indicating the preservative addition time for a plurality of blood samples.

[0069] A preservative addition time trained model is created by machine learning the correlation between these measurement data and preservative addition time data. The machine learning algorithm described above in the "Method for generating a disease trained model" section can be used as the algorithm used for machine learning.

[0070] In one embodiment, the preservative addition time trained model may be generated by qualitatively training a learning model on the appropriateness of the preservative addition time. For example, the preservative addition time trained model is generated using, as training data, first measurement data obtained from a blood sample to which an additive was added within an appropriate time and second measurement data obtained from a blood sample to which an additive was not added within an appropriate time. In this case, the preservative addition time trained model determines whether the preservative addition time is appropriate and outputs the appropriateness, thereby enabling the testing process to test whether the preservative addition time is appropriate.

[0071] In one aspect, the preservative addition time trained model may be generated by quantitatively training a learning model on the preservative addition time. For example, the preservative addition time is determined using the preservative addition time trained model, and the determination result of the preservative addition time is output, for example, as a judgment value. The judgment value here is not a binary value of 0 or 1, but a value having a predetermined numerical range. Then, in the inspection process, the appropriateness of the preservative addition time is inspected based on a comparison between the output judgment value and a threshold value. That is, in the inspection process, if it is determined that the output judgment value exceeds the threshold value, the inspection result indicates that the preservative addition time is inappropriate, and if it is determined that the output judgment value is equal to or less than the threshold value, the inspection result indicates that the preservative addition time is appropriate.

[0072] As described above, in this embodiment, a preservative addition time trained model is generated by having a learning model qualitatively or quantitatively learn whether the preservative addition time is appropriate. The preservative addition time trained model then determines whether the preservative addition time is appropriate and outputs the appropriateness, thereby testing the appropriateness of the preservative addition time in the testing process. Therefore, the testing process displays a test result indicating that the preservative addition time is appropriate or inappropriate. The preservative addition time trained model may be generated, for example, so that the preservative addition time is determined as two values, appropriate (good) or inappropriate (bad), or so that the preservative addition time is determined as three values, appropriate (good), warning, and inappropriate (bad).

[0073] <Disease Determination Method 2> In one embodiment of the present disclosure, the disease determination method is a method for determining the presence or absence of a disease in a subject using small RNA data in a blood sample for which the above-mentioned preservative addition time testing method has yielded a test result indicating that the preservative addition time is appropriate, and includes determining the presence or absence of the disease by inputting the small RNA data of the subject into a disease-trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data indicating the presence or absence of the disease in the multiple subjects. Hereinafter, the step of determining the presence or absence of the disease is also referred to as "disease determination step 2."

[0074] In the disease determination method according to this embodiment, the presence or absence of a disease is determined for a blood sample for which the test result indicating that the preservative addition time is appropriate is obtained by the above-described test method for preservative addition time. If the test result indicating that the preservative addition time is inappropriate is obtained by the above-described test method for preservative addition time, the determination of the presence or absence of a disease may or may not be performed. Note that, if the test result indicates that the preservative addition time is inappropriate, for example, the test result indicating that the preservative addition time is inappropriate may be presented to the user (i.e., the person performing the test method).

[0075] In one embodiment, it is preferable to determine a disease using only blood samples for which the test result indicates that the preservative addition time is appropriate, without using blood samples for which the test result indicates that the preservative addition time is inappropriate. In this case, if the test result indicates that the preservative addition time is inappropriate, the presence or absence of a disease is not determined. This can prevent erroneous determinations caused by the preservative addition time being inappropriate.

[0076] In one embodiment, even if a test result indicating that the preservative addition time is inappropriate is obtained, the disease determination step 2 may be executed, for example, to obtain a determination result as reference data. In this case, for example, after a test result indicating that the preservative addition time is inappropriate is obtained, the fact that a determination has been made may be presented to the user (i.e., the person executing the disease determination step 2).

[0077] When the disease determination step 2 is performed using only blood samples for which the test result indicates that the preservative addition time is appropriate, the disease determination step is performed after the testing step, but in one embodiment, the disease determination step may be performed before the testing step. In this case, if the test result in the testing step indicates that the preservative addition time is inappropriate, the determination result of the disease determination step 2 is treated as, for example, reference data.

[0078] Examples of the subject include the subject described above in the section "Disease Determination Method 1." Examples of the disease to be determined using the disease trained model include the diseases to be determined described above in the section "Disease Determination Method 1."

[0079] The disease trained model is generated by machine learning the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease incidence data showing the presence or absence of disease in the multiple subjects. The method for generating the disease trained model conforms to the method for generating the disease trained model described above in the section "Disease Determination Method 1."

[0080] <Disease Determination Model Generation Method> The disease determination model generation method according to one embodiment of the present disclosure is a disease determination model generation method that, when generating a disease determination model for determining a disease in a subject based on the expression levels of a plurality of small RNAs, selects a specimen from which a plurality of small RNAs used to generate the disease determination model are derived, wherein a specimen for which a preservative addition time between collection of the specimen from the subject and addition of the preservative is appropriate is selected as a standard specimen (also referred to as a "selection step" in the description of this embodiment), and the disease determination model is generated using the expression levels of the plurality of small RNAs in the standard specimen (also referred to as a "model generation step" in the description of this embodiment). According to the disease determination model generation method according to this embodiment, a specimen for which a preservative addition time is appropriate is selected as a standard specimen to create a model, thereby enabling the construction of a model in which noise caused by deterioration of the blood sample is removed.

[0081] Examples of the test subject include the test subjects described above in the section "Disease Determination Method 1." Examples of the disease to be determined using the disease determination model include the diseases to be determined described above in the section "Disease Determination Method 1."

[0082] The selection process is a process of selecting as a standard sample a sample for which the preservative addition time, from collection of the sample from the subject to addition of the preservative, is appropriate. Small RNAs are present in trace amounts in samples and are very unstable. Therefore, if the preservative addition time is inappropriate, the measurement results of the small RNA expression level may not accurately reflect the expression level of the small RNA contained in the sample. Specifically, as described above, if the preservative addition time exceeds a predetermined time (e.g., 4 hours), the measured expression level of the small RNA may change due to deterioration of the sample, etc. Therefore, in the selection process, a sample for which the preservative addition time is within a predetermined time (e.g., 4 hours) is selected as a standard sample.

[0083] The standard sample is a sample for which the preservative addition time is appropriate, and may be, for example, a sample prepared by the aforementioned method for preparing a blood sample for small RNA measurement, or a sample preserved by the aforementioned method for preserving a blood sample for small RNA measurement. The appropriate preservative addition time may be within 4 hours, preferably within 2 hours, and more preferably within 1 hour. In one aspect, the standard sample may be a sample corresponding to the small RNA data in a blood sample for which the test result indicating that the preservative addition time is appropriate was obtained by the aforementioned method for testing the preservative addition time.

[0084] The selection of standard specimens is performed, for example, by the creator of the disease determination model (i.e., the executor who executes the model generation method of this embodiment) managing the preservative addition time and selecting specimens whose preservative addition time is within a predetermined time (e.g., 4 hours). Thus, the selection of standard specimens can be performed, for example, by a method dependent on the creator, in which the creator manages the preservative addition time, but is not limited to this method. For example, the selection of standard specimens may be performed using a preservative addition time trained model. The details of the preservative addition time trained model are as described above.

[0085] The model generation step is a step of generating a disease determination model using the expression levels of multiple small RNAs in standard samples selected in the selection step. In the model generation step, for example, a disease determination model is generated to determine the presence or absence of a disease using a discrimination criterion based on the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in standard samples collected from multiple subjects and disease prevalence data showing the presence or absence of a disease in the multiple subjects. Examples of the discrimination criterion include the criteria described above in the "Disease Determination Method 1" section. In addition, in the model generation step, for example, a disease determination model is generated by machine learning the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in standard samples collected from multiple subjects and disease prevalence data showing the presence or absence of a disease in the multiple subjects. Details of the machine learning of the measurement data, disease prevalence data, and correlations can be applied as described above in the "Disease Trained Model Generation Method." However, in the measurement data (i.e., training data) to be machine-learned, "blood sample" is replaced with "standard sample."

[0086] In one embodiment, the disease determination model may be generated by having a learning model qualitatively learn the presence or absence of a disease. For example, a disease determination model is generated using, as training data, first measurement data obtained from a standard sample derived from a subject suffering from the disease and second measurement data obtained from a standard sample derived from a subject not suffering from the disease. In this case, the presence or absence of a disease is determined by inputting RNA data derived from the subject into the disease determination model. Therefore, in the disease determination step, a determination result of the presence or absence of a disease is displayed. When there are multiple types of diseases to be determined, a determination result of the presence or absence of a disease is displayed for each of the multiple types of diseases.

[0087] In one embodiment, the disease determination model may be generated by quantitatively learning a disease using a learning model. In this case, the disease learning model determines, for example, the probability of disease onset, and the determination result is output as a determination value. Then, the presence or absence of the disease is determined based on a comparison between the output determination value and a threshold value.

[0088] <Testing device, testing system, testing program, and recording medium> A testing device in one embodiment of the present disclosure is a testing device that tests the preservative addition time from blood collection to the addition of a preservative, using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in a blood sample, and includes a processor, wherein the processor tests the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time for the multiple blood samples.

[0089] A testing system in one embodiment of the present disclosure is a testing system that tests the preservative addition time from blood collection to the addition of a preservative using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in a blood sample, and includes: a measurement unit that obtains the small RNA data; and a testing unit that tests the preservative addition time by inputting the small RNA data obtained by the measurement unit into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time for the multiple blood samples.

[0090] A testing program in one embodiment of the present disclosure is a testing program for causing a computer to execute a testing process for testing the preservative addition time from blood collection to the addition of a preservative, using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in blood samples, by inputting the small RNA data into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time for the multiple blood samples.

[0091] In one embodiment of the present disclosure, the recording medium is a non-transitory recording medium having recorded thereon a testing program for causing a computer to execute a testing process for testing the preservative addition time from blood collection to the addition of a preservative using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in blood samples, by inputting the small RNA data into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time for the multiple blood samples.

[0092] Details of the inspection device, inspection system, inspection program, and inspection of the preservative addition time in the recording medium are as described above in the section "Method for inspecting preservative addition time." The program of the present disclosure can be provided as a program product. The program product includes any type of product for providing the program. For example, the program product includes a program provided via a network such as the Internet, and a non-transitory computer-readable recording medium such as a CD-ROM or DVD on which the program is stored.

[0093] An example of an inspection system and a measurement device according to an embodiment of the present disclosure will be described below with reference to FIGS. 1 and 2. FIG.

[0094] [Inspection System 20] As shown in FIG. 1, the inspection system 20 includes a measurement device 21 and an inspection device 30.

[0095] [Measurement Device 21] The measurement device 21 is a device that measures the expression levels of multiple small RNAs contained in a blood sample. As the measurement device 21, for example, an NGS is used.

[0096] [Testing Device 30] The testing device 30 is an example of a testing unit. The testing device 30 is a device that performs the testing process described above. That is, the testing device 30 acquires small RNA data indicating the results of measuring multiple small RNAs in a blood sample, and inputs the acquired small RNA data into a preservative addition time trained model, thereby testing the preservative addition time from blood collection to the addition of the preservative.

[0097] Furthermore, the testing device 30 may execute a disease determination step. In this case, the testing device 30 determines the presence or absence of a disease by inputting the small RNA data into a disease-trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data indicating the presence or absence of a disease in the multiple subjects.

[0098] 1, the inspection 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.

[0099] The CPU 31 (an example of a processor) is a central processing unit that executes various programs and controls each part. That is, the CPU 31 reads a program from the ROM 32 or the storage 34 and executes the program 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 program stored in the ROM 32 or the storage 34. The CPU 31 is an example of a processor.

[0100] 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.

[0101] In this embodiment, for example, an inspection program for executing an inspection process that performs the above-described inspection method is recorded in the storage 34. The inspection program may be a single program, or a group of programs configured from multiple programs or modules. The inspection 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.

[0102] 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.

[0103] 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 small RNAs measured by the measurement device 21 as input.

[0104] The display unit 36 ​​is, for example, a liquid crystal display, and displays various information. The testing device 30 can present the test results of the preservative addition time and the determination result of the presence or absence of a disease to the user through the display unit 36. The display unit 36 ​​may be a touch panel type and function as the input unit 35.

[0105] 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).

[0106] As shown in FIG. 2, in the testing device 30, the CPU 31 executes the testing program to function as a testing function unit 160 and a disease determination unit 170.

[0107] The inspection function unit 160 executes the inspection process. That is, in the inspection process, the inspection function unit 160 inspects the preservative addition time from blood collection to the addition of the preservative by inputting the small RNA data obtained by the measurement device 21 into the preservative addition time trained model.

[0108] The disease determination unit 170 executes a disease determination step. That is, the disease determination unit 170 determines the presence or absence of a disease by inputting the small RNA data obtained by the measurement device 21 into a disease-trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data indicating the presence or absence of a disease in the multiple subjects.

[0109] In this embodiment, the inspection system 20 includes the measuring device 21 and the inspection device 30, but the inspection system 20 may be configured with a single device. In this case, the single device functions as an example of a measurement unit and an inspection unit.

[0110] Furthermore, the testing device 30 may be composed of multiple devices. For example, the testing device 30 may be composed of multiple (e.g., two) devices that share the testing process and disease determination process. In this embodiment, the testing device 30 has the testing function unit 160 and the disease determination unit 170, but the testing device 30 does not necessarily have to have the disease determination unit 170.

[0111] Next, embodiments of the present disclosure will be described in detail using examples, but the embodiments of the present disclosure are not limited to these examples.

[0112] <Preservative Solution Formulation> The following two formulations (A) and (B) were used.

[0113] (Formulation (A)) Guanidine thioisocyanate: 4.3 M N-methylacetamide: 3.3 M Thioglycerin: 1% by volume Sodium acetate: 10 mM

[0114] (Formulation (B)) Guanidine thioisocyanate: 4.3 M Thioglycerin: 1% by volume Sodium acetate: 10 mM

[0115] <Preparation of Blood Samples> Blood was collected from the vein of each subject using a serum vacuum blood collection tube. After collection, the blood was left at room temperature (approximately 25°C) for at least 30 minutes to allow the coagulation components to coagulate, and then centrifuged at 2,000G for 10 minutes. 300 μL of serum was dispensed from the blood collection tube into a separate container. The entire process from blood collection to dispensing was completed within one hour. The sample was then left at room temperature (approximately 25°C). One hour, two hours, four hours, six hours, or 24 hours after collection, a preservative solution was added to the serum, and the sample was left at 25°C for seven days (also referred to as the 1-hour, 2-hour, 4-hour, 6-hour, or 24-hour sample, respectively). As a positive control, serum without the addition of preservative solution was stored in a deep freezer at -80°C within two hours of collection. As a negative control, 300 μL of serum without the addition of preservative solution was left at 25°C for seven days.

[0116] <RNA extraction and NGS measurement> Purified RNA solutions were obtained from the serum supplemented with preservative solution, the positive control, and the negative control. Next-generation sequencing (NGS) libraries were prepared by ligating 3' adapters, ligating 5' adapters, reverse transcription, PCR, and sequencing according to Seda Eminaga et al., Quantification of microRNA Expression with Next-Generation Sequencing, Curr Protoc Mol Biol. 2013 July; 04: Unit 4.17.

[0117] <Evaluation of the effect of adding preservative solution> (Formulation (A)) For the small RNAs (microRNAs) with the top 50 highest expression levels, the correlation between the expression level of the small RNA in the positive control and the expression levels of the small RNA in the samples taken 1 hour, 2 hours, 4 hours, 6 hours, and 24 hours later, as well as in the negative control, was evaluated. The obtained coefficient of determination R ２ are shown in Table 1. In the table, PC indicates a positive control, and NC indicates a negative control.

[0118]

[0119] (Formulation (B)) As with formulation (A), the correlation between the expression level of the small RNA in the positive control and the expression level of the small RNA in the samples 1 hour, 2 hours, 4 hours, 6 hours, and 24 hours later, and in the negative control, was evaluated for the top 50 small RNAs in terms of expression level. The obtained coefficient of determination R ２ are shown in Table 2. In the table, PC indicates a positive control, and NC indicates a negative control.

[0120]

[0121] In both formulations (A) and (B), the coefficient of determination R ２A good correlation was observed, with a coefficient of determination of 0.97 or higher. The correlation decreased in the 4-hour, 6-hour, and 24-hour samples, as well as the negative control, in that order. From the above, it was found that if the preservative was added within 2 hours of blood collection, the sample maintained a state almost identical to that of the positive control when stored at -80°C within 2 hours of collection. In the 4-hour sample, slight deviation from the positive control was observed due to degradation of small RNA at 25°C, but the coefficient of determination still exceeded 0.90, indicating only a minor change in state. However, in the 6-hour and 24-hour samples, degradation of small RNA progressed even at 25°C, resulting in a large difference from the positive control, with the coefficient of determination being less than 0.90. The coefficients of determination for the 6-hour and 24-hour samples were higher than that for the negative control (no preservative added), indicating the effect of the preservative, but degradation in the 6-hour sample was similar to that in the 24-hour sample, and the difference from the 4-hour sample was significant. That is, the later the timing of preservative addition, the more the initial small RNA expression distribution changes.

[0122] <Creation of a Preservative Addition Time Trained Model> A preservative addition time trained model I was created to distinguish between the following two groups. Based on the following conditions 1 to 4, the training data obtained under good conditions was designated as a good data group, and the training data obtained under poor conditions was designated as a poor data group. The preservative addition time trained model I was generated using training data for 112 samples. In other words, the preservative addition time trained model I is a model that makes a binary judgment of good or bad. The only information input when creating the training model is the expression levels of each of the multiple small RNAs in the multiple samples and whether the multiple samples were "good" or "poor." Information such as the preservative solution addition time, storage temperature, and storage period was not input. In this case, the training model was created using comprehensive small RNAs rather than specific small RNAs. A support vector machine algorithm was used for machine learning.

[0123] (Condition 1) [Good condition] Positive control, 1 hour later [Poor condition] Negative control, 2 hours, 4 hours, 6 hours, 24 hours later (Condition 2) [Good condition] Positive control, 1 hour, 2 hours later [Poor condition] Negative control, 4 hours, 6 hours, 24 hours later

[0124] (Condition 3) [Good condition] Positive control, 1 hour, 2 hours, 4 hours later [Poor condition] Negative control, 6 hours, 24 hours later

[0125] (Condition 4) [Good condition] Positive control, 1 hour, 2 hours, 4 hours, 6 hours later [Poor condition] Negative control, 24 hours later

[0126] <Obtaining Small RNA Data> Blood was collected from the subject, serum was separated, and samples were prepared with unknown conditions for the timing of adding preservative solution to the serum. The storage temperature was room temperature (approximately 25°C), and the storage period was within 7 days. The expression levels of small RNAs were comprehensively measured using NGS, and expression level data (small RNA data) showing the results of measuring multiple small RNAs were obtained.

[0127] <Verification of the model trained on preservative addition time> Small RNA data was input into the model trained on preservative addition time, and it was determined whether the data belonged to the good group or the bad group. The judgment performance was evaluated based on the accuracy rate of the good / bad judgment.

[0128] (Formulation (A)) The verification results using formulation (A) as the preservative solution are shown in the table below.

[0129]

[0130]

[0131]

[0132]

[0133] The specificity of condition 1 was 86% (12 / 14) and the sensitivity was 94% (33 / 35). The specificity of condition 2 was 100% (21 / 21) and the sensitivity was 100% (28 / 28). The specificity of condition 3 was 100% (28 / 28) and the sensitivity was 86% (18 / 21). The specificity of condition 4 was 100% (35 / 35) and the sensitivity was 100% (14 / 14).

[0134] (Formulation (B)) The verification results using formulation (B) as the preservative solution are shown in the table below.

[0135]

[0136]

[0137]

[0138] The specificity of condition 1 was 93% (13 / 14) and the sensitivity was 100% (35 / 35). The specificity of condition 2 was 95% (20 / 21) and the sensitivity was 93% (26 / 28). The specificity of condition 3 was 96% (27 / 28) and the sensitivity was 86% (18 / 21). The specificity of condition 4 was 100% (35 / 35) and the sensitivity was 100% (14 / 14).

[0139] The above results demonstrate that the preservative addition time trained model can detect whether the preservative solution was added properly. This provides information to reduce the risk of misdiagnosis when using samples to which the preservative solution was not added properly for disease discrimination.

[0140] <Positive rate in lung cancer discrimination model> The expression levels of small RNAs in 96 serum samples from lung cancer patients and healthy individuals, which had been centrifuged, serum fractionated, and frozen at -80°C within two hours of blood collection, were measured using NGS, and a lung cancer discrimination model was created by machine learning using the obtained small RNA data.

[0141] The small RNA data of the positive control (-80°C), 2-hour, 4-hour, 6-hour, and 24-hour samples, and the negative control (no preservative solution added) were input into this lung cancer discrimination model, and the positive detection rate (i.e., the probability of being erroneously identified as lung cancer) was examined. Specifically, a prediction score of 0.5 or higher by the lung cancer discrimination model was determined to be positive (Figure 3).

[0142] As shown in Table 11, there were almost no false positives in the positive control and 2-hour samples, while the false positive rate was 12.5% ​​in the 4-hour samples. The false positive rates were even higher in the 6-hour and 24-hour samples, as well as the negative control. These results demonstrate that a blood sample preparation method in which a preservative is added within 4 hours of blood collection is useful for accurately identifying lung cancer. Furthermore, adding a preservative within 4 hours achieved the same preservation effect as freezing at -80°C within 2 hours, even in a 25°C environment. Furthermore, detecting samples in which the preservative was added late or not added at all, demonstrated the potential for preventing false positives.

[0143] In the table below, the number of positive results indicates the number of samples determined to be lung cancer out of all samples derived from healthy individuals. The positive rate indicates the proportion (%) of all samples determined to be lung cancer.

[0144]

[0145] The disclosure of Japanese Patent Application No. 2023-182775, filed on October 24, 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.

[0146] 20 Testing system 21 Measuring device (an example of a measuring unit) 30 Testing device (an example of a testing unit) 31 CPU (an example of a processor) 32 ROM 33 RAM 34 Storage (an example of a recording medium) 35 Input unit 36 ​​Display unit 37 Communication interface 39 Bus 160 Testing function unit 170 Disease determination unit

Claims

1. A method for preparing a blood sample for measuring small RNA, comprising: separating serum or plasma from blood collected from a living body; and adding a preservative to the serum or plasma within 4 hours of collecting the blood.

2. The method of claim 1, wherein the time between drawing the blood and adding the preservative is within 2 hours.

3. The method of claim 1, wherein the time between drawing the blood and adding the preservative is within one hour.

4. The method of claim 1, wherein the preservative comprises a guanidinium salt.

5. The method of claim 1, wherein separating the serum or plasma comprises centrifuging the blood, and the time from collection of the blood to completing the centrifugation is within 2 hours.

6. The method of claim 5, wherein the time from collection of the blood to completion of the centrifugation is within one hour.

7. A method for storing a blood sample for measuring small RNA, comprising storing a blood sample prepared by the method according to any one of claims 1 to 6.

8. A method for measuring the expression levels of small RNAs, comprising: preparing a blood sample in which a preservative is added to serum or plasma separated from blood within 4 hours of collecting the blood; and measuring the expression levels of multiple small RNAs in the blood sample.

9. A method for diagnosing a disease in a subject using small RNA data indicating the expression levels of small RNAs obtained by the method described in claim 8, comprising determining the presence or absence of the disease using a discrimination criterion based on the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data indicating the presence or absence of the disease in the multiple subjects.

10. The method according to claim 9, wherein the presence or absence of a disease is determined by inputting small RNA data of the test subject into a disease-trained model that has been machine-learned to determine the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data showing the presence or absence of the disease in the multiple subjects.

11. A testing method for testing a preservative addition time from blood collection to addition of a preservative using small RNA data showing the results of measuring the expression levels of multiple small RNAs in a blood sample, the testing method comprising: testing the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has machine-learned the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition time in the multiple blood samples.

12. A method for generating a preservative addition time trained model used in a testing method for testing a preservative addition time from blood collection to addition of a preservative, the testing method comprising: inputting small RNA data indicating the results of measuring the expression levels of multiple small RNAs in blood samples into a preservative addition time trained model; the method for generating a preservative addition time trained model includes machine learning the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition times in the multiple blood samples, to generate the preservative addition time trained model.

13. A method for diagnosing a disease in a subject using small RNA data in a blood sample for which a test result indicating that the preservative addition time is appropriate has been obtained by the test method described in claim 11, the method comprising: determining the presence or absence of the disease by inputting the small RNA data of the subject into a disease-trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in blood samples collected from multiple subjects and disease prevalence data indicating the presence or absence of the disease in the multiple subjects.

14. A method for generating a disease determination model for determining a disease in a subject based on the expression levels of multiple small RNAs, comprising the steps of: selecting a specimen from which multiple small RNAs used in generating the disease determination model are derived, wherein a specimen in which the preservative addition time between collection of the specimen from the subject and addition of the preservative is appropriate is selected as a standard specimen; and the disease determination model is generated using the expression levels of the multiple small RNAs in the standard specimen.

15. The method according to claim 14, wherein the standard specimen is a specimen corresponding to small RNA data in a blood sample for which the test result indicating that the preservative addition time is appropriate is obtained by the test method described in claim 11.

16. A testing device that tests a preservative addition time from blood collection to addition of a preservative using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in a blood sample, the testing device having a processor, wherein the processor tests the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has machine-learned the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time in the multiple blood samples.

17. A testing system that tests a preservative addition time from blood collection to addition of a preservative using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in a blood sample, comprising: a measurement unit that obtains the small RNA data; and a testing unit that tests the preservative addition time by inputting the small RNA data obtained by the measurement unit into a preservative addition time trained model that has been machine-learned to determine the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time in the multiple blood samples.

18. A testing program for causing a computer to execute a testing process for testing a preservative addition time from blood collection to addition of a preservative using small RNA data indicating the results of measuring the expression levels of multiple small RNAs in blood samples, the testing process testing the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has machine-learned the correlation between measurement data indicating the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data indicating the preservative addition time in the multiple blood samples.

19. A non-temporary recording medium having recorded thereon a testing program for causing a computer to execute a testing process for testing the preservative addition time from blood collection to the addition of a preservative using small RNA data showing the results of measuring the expression levels of multiple small RNAs in blood samples, the testing process testing the preservative addition time by inputting the small RNA data into a preservative addition time trained model that has machine-learned the correlation between measurement data showing the results of measuring the expression levels of multiple small RNAs in multiple blood samples and preservative addition time data showing the preservative addition time in the multiple blood samples.

Images