Microsatellite instability detection method independent of contrast and reference baselines
By screening differential and constant stable sites in samples of the same cancer type, sequencing and Pearson correlation analysis, the problem of control and reference baseline dependence in the prior art was solved, and efficient and accurate microsatellite instability was achieved. Testing reduces the cost of testing.
Patent Information
- Application Number
- CN202411283765.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-13
- Publication Date
- 2024-12-31
AI Technical Summary
The existing microsatellite instability detection methods rely on control and reference baselines, resulting in limited collection of negative samples, high detection costs, low accuracy, and insufficient consideration of the instability differences in sites in different cancer samples.
Using a method that does not rely on control and reference baseline, differential sites and constant stable sites are screened through samples of the same cancer type, sequencing and Pearson correlation analysis are carried out, reference dictionary is constructed, and the stability judgment of microsatellite sites is achieved and the stability of microsatellite sites is reduced. Inspection cost and improve accuracy.
It solves the problems of low detection accuracy and high cost, improves the sensitivity and accuracy of microsatellite instability detection, reduces the dependence on negative samples, and simplifies the sample collection process.
Smart Images
Figure CN119229961A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of microsatellite instability detection, and in particular to a microsatellite instability detection method that is independent of control and reference baseline. Background Art
[0002] Microsatellite is a short tandem repeat sequence widely distributed in the genome of eukaryotes, with 1 to 6 nucleotides as the repeating unit. Microsatellite instability (MSI) refers to the phenomenon that insertion or deletion mutations occur during DNA replication, causing changes in the length of the microsatellite locus sequence, which is related to the body's DNA mismatch repair defects. MSI is closely related to the occurrence and development of many cancers.
[0003] Studies in recent years have also shown that tumor patients with different MSI status can benefit differently from immunotherapy. For example, immunotherapy for metastatic colorectal cancer is only effective for mismatch repair-deficient tumors with high microsatellite instability. The 2020 NCCN guidelines clearly state that "pembrolizumab is recommended for the second-line treatment of advanced endometrial cancer with MSI-H (microsatellite instability high) / dMMR (defective mismatch repair) and advanced cervical cancer with positive PD-L1 expression or MSI-H / dMMR". Therefore, with the widespread use of immunotherapy in tumor treatment, accurate detection of a patient's MSI status is of great value for disease classification, clinical treatment methods, and prognosis prediction.
[0004] Some conventional NGS-based MSI detection methods require tumor samples and negative control samples. For example, MSIsensor directly compares the read length distribution between tumor samples and healthy samples based on the chi-square test to determine the mutation status of the site. Another part of the method requires a large number of healthy human samples to construct some statistical reference baselines. For example, MSIsensor-pro first uses normal human samples to model and infer the probability of insertion and deletion of each microsatellite site during normal DNA replication as the reference baseline of the site, and then determines whether the probability of insertion or deletion of the sample to be tested is significantly different from that of normal people to determine whether the site is mutated. mSINGS also first establishes the mean and standard deviation of the number of repetitions of each site through negative control samples or MSS tumor samples as a reference baseline, and then determines whether the number of repetitions of the site in the sample to be tested exceeds the (mean + 3*standard deviation) obtained by statistics of the site in the negative control or MSS tumor samples to determine the stability of the site.
[0005] However, negative control samples are not available in some scenarios, which limits the use of methods that require tumor samples and negative control samples. The method of establishing a reference baseline requires more samples from healthy people. The difficulty of collecting and the high cost of sequencing also limit the use of this method. At the same time, most of the existing MSI detection methods do not perform a first-step screening of the sites, or do not fully consider the differences in the frequency of instability of the same site in different cancer samples when screening the sites. Summary of the invention
[0006] In view of the above-mentioned deficiencies in the prior art, the present invention provides a microsatellite instability detection method that is independent of controls and reference baselines, which solves the problems of high detection costs and low accuracy in the prior art due to limited negative sample collection and inaccurate site screening.
[0007] In order to achieve the above-mentioned purpose of the invention, the technical solution adopted by the present invention is: a method for detecting microsatellite instability that does not rely on controls and reference baselines, comprising the following steps:
[0008] S1: The cancer type of the sample to be tested is known, and the site screening is performed based on samples of the same cancer type to obtain a differential site set suitable for microsatellite instability detection;
[0009] S2: The cancer type of the sample to be tested is known, and a set of constant stable sites that are always stable and have a detection rate that meets the requirements are screened based on samples of the same cancer type;
[0010] S3: Sequence the differential loci and constant stable loci of the sample to be tested, and extract the distribution of the number of repetitions of microsatellite loci;
[0011] S4: Based on the distribution of the number of repetitions, the microsatellite frequency distribution is extracted for the constant stable site set and a reference dictionary is constructed. For the differential site set, the data in the reference dictionary are extracted according to its repetition unit and Pearson correlation analysis is performed to determine the stability of the site.
[0012] S5: Obtain the number of unstable sites of the sample to be tested, and further obtain the instability score of the current sample;
[0013] S6: Based on the instability score of the current sample, the microsatellite instability of the current sample is determined, and the microsatellite instability detection based on a single tumor is completed.
[0014] The beneficial effect of the above scheme is that the microsatellite instability detection method provided by the present invention refers to the distribution of the number of repetitions of those conservatively stable sites in the same cancer type to judge the stability of the effective detection sites obtained by the microsatellite site screening method, thereby performing microsatellite instability detection on the test sample, solving the problems of low detection accuracy and high detection cost caused by limited negative sample collection in the existing method.
[0015] Furthermore, the site screening comprises the following steps:
[0016] A1: Collect paired microsatellite instability positive and negative samples of tumor and adjacent adjacent tissues according to cancer type;
[0017] A2: Perform data preprocessing, comparison and instability detection on microsatellite instability positive and negative samples, and count the test results of each sample, and remove sites in the samples with a detection ratio lower than the set threshold;
[0018] A3: Hypothesis testing method was used to test the difference in the proportion of instability of the retained loci in microsatellite instability-positive samples and negative samples;
[0019] A4: Remove the sites that the hypothesis testing method believes have no significant difference in the proportion of instability between microsatellite instability-positive samples and negative samples, obtain a set of effective sites that support the detection of microsatellite instability in the cancer type, and achieve site screening.
[0020] The beneficial effect of the above further scheme is that the microsatellite site screening method of the scheme is used to select sites that can effectively infer the state of microsatellite instability based on the sequencing data of a specific cancer sample, and provide them to the microsatellite instability detection method as a detection site set. The present invention fully considers the differences in the proportion of instability occurring in different microsatellite sites, and also considers the differences in the instability occurring in sites in different tumors, uses statistical methods to test microsatellite sites, screens site combinations suitable for samples, and improves the sensitivity of detection.
[0021] Furthermore, the hypothesis testing method in A3 is Fisher's exact test, and the calculation formula of Fisher's exact test is:
[0022]
[0023] Among them, a is the number of stable samples in the MSI sample, b is the number of stable samples in the MSS sample, c is the number of unstable samples in the MSI sample, and d is the number of unstable samples in the MSS sample.
[0024] The beneficial effect of the above further scheme is: by using the above Fisher's exact test to test whether there is a significant difference in the proportion of unstable changes in each site in microsatellite instability positive samples and negative samples.
[0025] Furthermore, constructing the reference dictionary in S4 includes the following steps:
[0026] S41: convert the repetition number distribution of each site in the constant stable site set into a repetition number frequency distribution;
[0027] S42: Shift the frequency distribution of the number of repetitions, align the highest peak as the center, use the site repeat unit as the key of the dictionary, and use the shifted frequency distribution of the number of repetitions as the value to construct a reference dictionary.
[0028] Furthermore, in S4, the Pearson correlation analysis is performed on the data in the reference dictionary extracted from the differential site set according to its repeating units, including the following sub-steps:
[0029] S43: converting the repetition number distribution of each site in the differential site set into a repetition number frequency distribution;
[0030] S44: Shift the frequency distribution of the number of repetitions, align the highest peak as the center, and take out all the values in the reference dictionary as templates according to the site repetition unit;
[0031] S45: Perform Pearson correlation analysis on each frequency distribution of the number of repetitions in the template and the frequency distribution of the number of repetitions after the current site to be detected is shifted.
[0032] The beneficial effect of the above further scheme is: by performing Pearson correlation analysis on the frequency distribution of each repetition number in the template and the frequency distribution of the repetition number after the current site to be detected is translated, the stability of the current site is judged.
[0033] Furthermore, the calculation formula of the Pearson correlation analysis is:
[0034]
[0035] Among them, ρ X,Y is the Pearson correlation coefficient, X is the frequency distribution of a repetition in the template, Y is the frequency distribution of the repetition after the current site to be detected is translated, i is the loop variable, n is the dimension of X and Y, x i is the i-th element in the X vector, y i is the i-th element in the Y vector, and E(·) is the sample mean.
[0036] Further, based on the Pearson correlation analysis, when the Pearson correlation coefficient ρ X,YIf the number greater than 0.9 exceeds the set ratio of the number of templates, the current site to be detected is a stable site.
[0037] Furthermore, the calculation formula for the sample instability score in S5 is:
[0038]
[0039] Among them, x is the number of instability and t is the total set of judgment sites.
[0040] The beneficial effect of the above further scheme is: by dividing the number of unstable sites by the number of sites in the judgment site set, the instability score of the current sample is obtained. When the instability score is greater than a certain threshold, the current sample is considered unstable. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 Flow chart of a method for detecting microsatellite instability that is independent of controls and reference baselines.
[0042] Figure 2 This is a heat map of the proportion of unstable sites in samples of various types of cancer.
[0043] Figure 3 Box plot of microsatellite instability scores of samples calculated using the screened differential sites and all sites.
[0044] Figure 4 This is the box distribution diagram of microsatellite instability scores of 6 cancer samples. DETAILED DESCRIPTION
[0045] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.
[0046] like Figure 1 As shown, a method for detecting microsatellite instability that does not rely on controls and reference baselines comprises the following steps:
[0047] S1: The cancer type of the sample to be tested is known, and the site screening is performed based on samples of the same cancer type to obtain a differential site set suitable for microsatellite instability detection;
[0048] S2: The cancer type of the sample to be tested is known, and a set of constant stable sites that are always stable and have a detection rate that meets the requirements are screened based on samples of the same cancer type;
[0049] S3: Sequence the differential loci and constant stable loci of the sample to be tested, and extract the distribution of the number of repetitions of microsatellite loci;
[0050] Differential site set: For a certain cancer, a set of sites with significantly different instability ratios between microsatellite instability-positive samples and negative samples is called a differential site set.
[0051] Constant stable site set: For a certain type of cancer, the set of sites that remain stable in both microsatellite instability positive and negative samples and whose detection ratios meet certain requirements is called the constant stable site set.
[0052] In this embodiment, only the differential site set and the constant stable microsatellite positions of the sequenced samples to be tested can be analyzed;
[0053] S4: Based on the distribution of the number of repetitions, the microsatellite frequency distribution is extracted for the constant stable site set and a reference dictionary is constructed. For the differential site set, the data in the reference dictionary are extracted according to its repetition unit and Pearson correlation analysis is performed to determine the stability of the site.
[0054] S5: Obtain the number of unstable sites of the sample to be tested, and further obtain the instability score of the current sample;
[0055] S6: Based on the instability score of the current sample, the microsatellite instability of the current sample is determined, and the microsatellite instability detection based on a single tumor is completed.
[0056] The site screening comprises the following steps:
[0057] A1: Collect paired microsatellite instability positive and negative samples of tumor and adjacent adjacent tissues according to cancer type;
[0058] A2: Perform data preprocessing, comparison and instability detection on microsatellite instability positive and negative samples, and count the test results of each sample, and remove sites in the samples with a detection ratio lower than the set threshold;
[0059] In this embodiment, some existing tools are used to perform data preprocessing, alignment, and microsatellite instability detection on the above tumor and adjacent adjacent paired samples; data preprocessing is to remove low-quality sequencing reads and duplicate reads caused by PCR, and microsatellite instability detection is to obtain microsatellite loci, repeat units, and locus states in the samples;
[0060] A3: Hypothesis testing method was used to test the difference in the proportion of instability of the retained loci in microsatellite instability-positive samples and negative samples;
[0061] A4: Remove the sites that the hypothesis testing method believes have no significant difference in the proportion of instability between microsatellite instability-positive samples and negative samples, obtain a set of effective sites that support the detection of microsatellite instability in the cancer type, and achieve site screening.
[0062] The hypothesis testing method in A3 is Fisher's exact test, and the calculation formula of Fisher's exact test is:
[0063]
[0064] Among them, a is the number of stable sites in MSI samples, b is the number of stable sites in MSS samples, c is the number of unstable sites in MSI samples, and d is the number of unstable sites in MSS samples. Sites with p-value less than 0.05 were retained, and it was considered that there was a significant difference in the proportion of instability in these sites in microsatellite instability-positive samples and negative samples.
[0065] In S3, the sequencing data is preprocessed and aligned, and the number of repetitions of microsatellite loci is extracted using some existing tools, such as the scan command of msisensor-pro.
[0066] The construction of the reference dictionary in S4 includes the following steps:
[0067] S41: convert the repetition number distribution of each site in the constant stable site set into a repetition number frequency distribution;
[0068] S42: Shift the frequency distribution of the number of repetitions, align the highest peak as the center, use the site repeat unit as the key of the dictionary, and use the shifted frequency distribution of the number of repetitions as the value to construct a reference dictionary.
[0069] In S4, the Pearson correlation analysis is performed on the data in the reference dictionary extracted from the differential site set according to its repeating units, including the following sub-steps:
[0070] S43: converting the repetition number distribution of each site in the differential site set into a repetition number frequency distribution;
[0071] S44: Shift the frequency distribution of the number of repetitions, align the highest peak as the center, and take out all the values in the reference dictionary as templates according to the site repetition unit;
[0072] S45: Perform Pearson correlation analysis on each frequency distribution of the number of repetitions in the template and the frequency distribution of the number of repetitions after the current site to be detected is shifted.
[0073] The calculation formula of the Pearson correlation analysis is:
[0074]
[0075] Among them, ρ X,Yis the Pearson correlation coefficient, X is the frequency distribution of a repetition in the template, Y is the frequency distribution of the repetition after the current site to be detected is translated, i is the loop variable, n is the dimension of X and Y, x i is the i-th element in the X vector, y i is the i-th element in the Y vector, and E(·) is the sample mean.
[0076] Based on the Pearson correlation analysis, when the Pearson correlation coefficient ρ X,Y If the number greater than 0.9 exceeds the set ratio of the number of templates, the current site to be detected is a stable site.
[0077] The calculation formula for the sample instability score in S5 is:
[0078]
[0079] Among them, x is the number of instability and t is the total set of judgment sites.
[0080] The proportion of unstable microsatellite loci in tumors varies, and the proportion of unstable microsatellite loci in different tumors also varies. Therefore, if microsatellite loci are not screened, the MSI detection effect will be greatly affected. Using such a locus set may make it difficult to detect some MSI patients, affecting the patients' benefits from some immunotherapy. Figure 2 As shown in the figure, 429, 441, and 532 samples of colon adenocarcinoma (COAD), gastric adenocarcinoma (STAD), and endometrial carcinoma (UCEC) were downloaded from TCGA, of which 72, 85, and 168 were microsatellite instability-positive samples, respectively. The sites detected in at least 30% of the samples were selected, and the proportion of instability in various types of samples was counted. It can be clearly seen that the proportion of instability in some sites in different types of tumors is quite different.
[0081] In one embodiment of the present invention, microsatellite locus screening is first performed:
[0082] Cancer samples in the Cancer Genome Atlas (TCGA) database were selected for site selection. Taking gastric adenocarcinoma as an example, whole exome sequencing data of 441 samples with microsatellite instability were obtained, including 85 microsatellite instability positive samples and 356 negative samples. The original data quality was analyzed, compared with the reference genome, duplicate reads were removed, and after data quality control, microsatellite instability detection was performed using msisensor to obtain microsatellite sites, repeat units, and site status in the samples;
[0083] According to the detection results of each sample site, a site stability state-sample table is established. Each row in the table is a site, each column is a sample, and the table content is the stability state of a site in the sample, marked as stable, unstable, and not detected;
[0084] The sites measured in more than 30% of the total samples are retained, and the number of stable sites in MSS samples, the number of unstable sites in MSS samples, the number of stable sites in MSI samples, and the number of unstable sites in MSI samples at the sites are counted (MSI-L and MSS samples are collectively referred to as MSS samples, and MSI-H samples are referred to as MSI samples). Assume that the statistical number of a site is as shown in Table 1:
[0085] Table 1 Site statistics table
[0086]
[0087]
[0088] Fisher's exact test was used to determine whether there was a significant difference in the proportion of unstable changes in the site in MSI samples and in MSS samples. If the p value was less than 0.05, it was considered to be significantly different and the site was added to the differential site set. Finally, a total of 3789 sites were obtained.
[0089] like Figure 3 As shown in the figure, the box plot shows the comparison of the instability scores of colon adenocarcinoma, gastric adenocarcinoma, and endometrial cancer samples based on all detected sites and calculated after site screening. After using msisensor to detect microsatellite instability in these samples, the microsatellite instability score is recorded as the number of instabilities in the detected sites divided by the total number of detected sites. It can be seen that after site screening, the difference in instability scores between microsatellite instability positive samples and negative samples has widened. Taking colon adenocarcinoma as an example, before site screening, the average instability score of positive samples was about 0.2, and after site screening, the average instability score of positive samples was about 0.5, and the instability score of negative samples did not change much, which means that it is easier to detect microsatellite instability positive samples.
[0090] Next, the sample microsatellite instability test is performed, relying only on the site stability state-sample table established above, including the following steps:
[0091] In the constructed site stability state-sample table, those sites that are detected in more than 30% of the samples and are stable in all the detected samples are screened, which are called site set constant stability;
[0092] For the tumor samples to be tested, the MSI method of MSISensor was used to obtain the distribution of the number of repeats of all microsatellite loci detected;
[0093] For the sites detected in the tumor sample to be tested and in a constant state, the repetition frequency distribution is converted into a repetition frequency distribution, and the repetition frequency distribution is translated to align the highest peak as the center, and the site repetition unit is used as the dictionary key and the translated distribution is used as the value to construct a reference dictionary, such as dict['A'] = [f1,f2,f3…f m ], that is, m stable sites are collected as references for the repeating unit A.
[0094] The sites detected in the tumor sample to be tested and in the differential site set are taken as the judgment site set, the site repetition frequency distribution is converted into the repetition frequency frequency distribution, and the distribution is translated with the highest peak as the center, and then all the values in the reference dictionary are taken as templates according to its repetition unit. For example, the repetition unit of a site to be judged is A, and its repetition frequency distribution after translation is recorded as f', then we can get m reference frequency distributions from the above reference dictionary, and put f1,f2,f3...f m Take them out one by one and calculate the Pierre's correlation coefficient with f'. If the number of times with high correlation exceeds a certain proportion of m, the current site is judged to be stable;
[0095] The number of unstable sites is divided by the number of sites in the judgment site set to obtain the instability score of the current sample; if the instability score is greater than a certain threshold, the current sample is considered unstable.
[0096] The present invention fully considers the differences in the instability of sites in different tumors, uses statistical methods to test microsatellite sites, and provides a screening method for sites suitable for detecting the stability state of sample microsatellites. It can be used to guide site selection in clinical use or scientific research, so that only the screened sites can be sequenced, which has certain economic benefits and is based on Figure 3 It is shown that the microsatellite loci screened using this method can more easily detect microsatellite instability-positive samples than before screening.
[0097] The present invention provides a method for detecting microsatellite instability, which does not require paired normal tissue samples as controls, nor does it require a large number of healthy human samples to construct a reference baseline. Microsatellite instability detection can be achieved using only tumor samples, which is simple and convenient. The test was conducted on 1402 samples of colon adenocarcinoma, gastric adenocarcinoma, and endometrial cancer downloaded from TCGA. The results are as follows Figure 4 As shown, there are significant differences in the instability scores of microsatellite instability-positive and -negative samples for each cancer. The AUCs of the three cancer samples were calculated to be above 99%, which means that the vast majority of samples can be classified correctly.
[0098] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the invention.
Claims
1. A method for detecting microsatellite instability that is independent of controls and reference baselines, characterized in that: The following steps are involved: S1: The cancer type of the sample to be tested is known, and the site screening is performed based on samples of the same cancer type to obtain a differential site set suitable for microsatellite instability detection; S2: The cancer type of the sample to be tested is known, and a set of constant stable sites that are always stable and have a detection rate that meets the requirements are screened based on samples of the same cancer type; S3: Sequence the differential loci and constant stable loci of the sample to be tested, and extract the distribution of the number of repetitions of microsatellite loci; S4: Based on the distribution of the number of repetitions, the microsatellite frequency distribution is extracted for the constant stable site set and a reference dictionary is constructed. For the differential site set, the data in the reference dictionary are extracted according to its repetition unit and Pearson correlation analysis is performed to determine the stability of the site. S5: Obtain the number of unstable sites of the sample to be tested, and further obtain the instability score of the current sample; S6: Based on the instability score of the current sample, determine the microsatellite instability of the current sample and complete the microsatellite instability detection that is independent of the control and reference baseline.
2. The method for detecting microsatellite instability independent of control and reference baseline according to claim 1, characterized in that: The site screening comprises the following steps: A1: Collect paired microsatellite instability positive and negative samples of tumor and adjacent adjacent tissues according to cancer type; A2: Perform data preprocessing, comparison and instability detection on microsatellite instability positive and negative samples, and count the test results of each sample, and remove sites in the samples with a detection ratio lower than the set threshold; A3: Hypothesis testing method was used to test the difference in the proportion of instability of the retained loci in microsatellite instability-positive samples and negative samples; A4: Remove the sites that the hypothesis testing method believes have no significant difference in the proportion of instability between microsatellite instability-positive samples and negative samples, obtain a set of effective sites that support the detection of microsatellite instability in the cancer type, and achieve site screening.
3. The method for detecting microsatellite instability independent of control and reference baseline according to claim 2, characterized in that: The hypothesis testing method in A3 is Fisher's exact test, and the calculation formula of Fisher's exact test is: Among them, a is the number of stable samples in the MSI sample, b is the number of stable samples in the MSS sample, c is the number of unstable samples in the MSI sample, and d is the number of unstable samples in the MSS sample.
4. The method for detecting microsatellite instability independent of control and reference baseline according to claim 1, characterized in that: The construction of the reference dictionary in S4 includes the following steps: S41: convert the repetition number distribution of each site in the constant stable site set into a repetition number frequency distribution; S42: Shift the frequency distribution of the number of repetitions, align the highest peak as the center, use the site repeat unit as the key of the dictionary, and use the shifted frequency distribution of the number of repetitions as the value to construct a reference dictionary.
5. The method for detecting microsatellite instability independent of control and reference baseline according to claim 4, characterized in that: In S4, the Pearson correlation analysis is performed on the data in the reference dictionary extracted from the differential site set according to its repeating units, including the following sub-steps: S43: converting the repetition number distribution of each site in the differential site set into a repetition number frequency distribution; S44: Shift the frequency distribution of the number of repetitions, align the highest peak as the center, and take out all the values in the reference dictionary as templates according to the site repetition unit; S45: Perform Pearson correlation analysis on each frequency distribution of the number of repetitions in the template and the frequency distribution of the number of repetitions after the current site to be detected is shifted.
6. The method for detecting microsatellite instability independent of control and reference baseline according to claim 5, characterized in that: The calculation formula of the Pearson correlation analysis is: Among them, ρ X,Y is the Pearson correlation coefficient, X is the frequency distribution of a repetition in the template, Y is the frequency distribution of the repetition after the current site to be detected is translated, i is the loop variable, n is the dimension of X and Y, x i is the i-th element in the X vector, y i is the i-th element in the Y vector, and E(·) is the sample mean.
7. The method for detecting microsatellite instability independent of control and reference baseline according to claim 6, characterized in that: Based on the Pearson correlation analysis, when the Pearson correlation coefficient ρ X,Y If the number greater than 0.9 exceeds the set ratio of the number of templates, the current site to be detected is a stable site.
8. The method for detecting microsatellite instability independent of control and reference baseline according to claim 1, characterized in that: The calculation formula for the sample instability score in S5 is: Among them, x is the number of instability and t is the total set of judgment sites.