A cancer diagnosis model based on free DNA and its application

Abstract

The present invention discloses a cancer diagnosis model and application based on free DNA. In the first aspect of the present application, a method for constructing a cancer diagnosis model is provided, and the construction method comprises the following steps: obtaining sequencing data of free DNA of non-cancer population; comparing the sequencing data of free DNA with the reference genome to obtain the corresponding sites of the ends of free DNA on the reference genome; constructing a cancer diagnosis model according to the corresponding sites of the ends of free DNA on the target genome interval. The present application adopts an approach opposite to the mainstream method. Instead of trying to identify diagnostic markers based on the specific fragmentation pattern of tumor cfDNA, a cancer diagnosis model is constructed by finding the terminal distribution pattern of cfDNA in non-cancer population. In this way, patients with various different cancers can be distinguished from the control population, with good specificity and sensitivity, and pan-cancer cancer diagnosis is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of cancer diagnosis, and in particular to a cancer diagnosis model and application based on free DNA. Background Art

[0002] Circulating cell-free DNA (cfDNA) in peripheral blood is a DNA fragment produced by natural degradation after cell death. Studies have shown that in healthy people, the main sources of cfDNA are blood cells and liver; in cancer patients, dead tumor cells also release cfDNA into the peripheral blood. There are many differences between cfDNA fragments from tumors and cfDNA fragments from non-tumor cells, such as gene mutations, chromosome copy number variations, abnormal methylation levels, and so on. Therefore, cancer diagnosis can be performed by detecting cfDNA signals from tumors in peripheral blood, that is, liquid biopsy technology. Since the collection of peripheral blood is almost non-invasive to the human body and can be sampled repeatedly, cancer diagnosis technology based on peripheral blood cfDNA has a high application value. Related studies have also identified a large number of tumor cfDNA markers and developed a large number of cancer diagnosis technologies, such as high-frequency gene mutation kits, tumor-specific methylation marker kits, and so on.

[0003] In recent years, researchers have discovered that cfDNA is not randomly fragmented when it is produced, and based on this principle, some tumor markers related to cfDNA fragmentation patterns have been discovered, and multiple cancer diagnostic technologies have been developed. For example, patients with liver cancer have a set of preferential cfDNA terminal sites distributed on the genome. These preferential terminal sites can distinguish liver cancer patients from the control group and can therefore be used as tumor markers. However, these preferential sites need to be identified using complex statistical methods. Moreover, each type of cancer needs to be identified separately, which is costly and has poor universality. The accuracy of diagnosis based on cfDNA length and terminal base usage patterns is too low, and the sensitivity is very limited in early cancer patients. Therefore, it is necessary to provide a cfDNA-based cancer diagnosis model with strong universality and high accuracy. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes a cfDNA-based cancer diagnosis model and application with strong universality and high accuracy.

[0005] In a first aspect of the present application, a method for constructing a cancer diagnosis model is provided, the method comprising the following steps:

[0006] Obtain sequencing data of cell-free DNA from non-cancer populations;

[0007] Compare the sequencing data of the free DNA with the reference genome to obtain the sites on the reference genome corresponding to the ends of the free DNA;

[0008] A cancer diagnosis model is constructed based on the corresponding sites of the ends of free DNA in the target genomic interval.

[0009] The construction method according to the embodiment of the present application has at least the following beneficial effects:

[0010] In the peripheral blood of non-cancer populations, cfDNA usually comes from normal cells, and the overall fragmentation pattern is relatively similar and highly consistent. In cancer patients, due to the high heterogeneity of the tumor itself, the fragmentation patterns of cfDNA released by tumors of different cancer types, or even different patients of the same cancer type, may be very different, resulting in the general dilemma of low universality and low sensitivity of tumor markers identified based on tumor patient samples. From another perspective, the applicant believes that although the cfDNA fragmentation patterns of each tumor patient may be different, they all have one thing in common, that is, they are all different from the DNA fragmentation patterns of non-cancer populations. Based on this principle, this application adopts an approach opposite to the mainstream method. Instead of trying to identify diagnostic markers based on tumor cfDNA-specific fragmentation patterns, it constructs a cancer diagnosis model by finding the terminal distribution pattern of cfDNA in non-cancer populations. In this way, patients with various different cancers can be distinguished from the control population, with good specificity and sensitivity, and pan-cancer cancer diagnosis is achieved.

[0011] In some embodiments of the present application, the end of the free DNA includes at least one of an upstream end and a downstream end.

[0012] In some embodiments of the present application, the reference genome is the entire human genome or a set interval of the entire human genome, or it may be the genome of one or more human individuals.

[0013] In some embodiments of the present application, the target genomic interval is the whole human genome or a set interval of the whole human genome.

[0014] In some embodiments of the present application, the target genomic interval is at least one of chr1, chr2, chr3, chr4, chr5, chr6, chr7, chr8, chr9, chr10, chr11, chr12, chr13, chr14, chr15, chr16, chr17, chr18, chr19, chr20, chr21, chr22, chrX, and chrY in the whole human genome.

[0015] Among them, the reference genome and the target genome interval can be the same or different, but obviously have overlapping parts.

[0016] In some embodiments of the present application, the method for constructing a cancer diagnosis model according to the corresponding sites of the ends of free DNA in the target genomic interval includes:

[0017] The number of times different sites in the target genomic interval appear as the ends of free DNA is calculated, and an end distribution model that characterizes the site-frequency relationship is constructed, which is the cancer diagnosis model.

[0018] A second aspect of the present application provides a sample classification method, the classification method comprising the following steps:

[0019] Obtain free DNA end distribution data of samples;

[0020] Compare the cell-free DNA end distribution data of the sample with the cell-free DNA end distribution data of a non-cancer population;

[0021] The classification of the sample is determined based on the comparison results.

[0022] In some embodiments of the present application, the method for obtaining free DNA end distribution data of a sample includes:

[0023] Obtain sequencing data of free DNA of the sample;

[0024] Compare the sequencing data of the sample's free DNA with the reference genome to obtain the corresponding sites of the ends of the sample's free DNA on the reference genome;

[0025] The number of times different sites in the target genomic interval appear as the ends of the free DNA of the sample is calculated.

[0026] In some embodiments of the present application, the free DNA end distribution data of non-cancer patients is obtained by the following method:

[0027] Obtain sequencing data of cell-free DNA from non-cancer populations;

[0028] The sequencing data of the free DNA of the non-cancer population is compared with the reference genome to obtain the corresponding sites of the ends of the free DNA of the non-cancer population on the reference genome;

[0029] The number of times different sites in the target genomic interval appear as the ends of the free DNA of the non-cancer population is calculated, and a terminal distribution model that characterizes the site-frequency relationship is constructed.

[0030] In some embodiments of the present application, the method for comparison includes comparing the consistency of the number of times different sites in the target genomic interval appear as the ends of the free DNA of the sample and the number of times they appear as the ends of the free DNA of a non-cancer population:

[0031] The consistency comparison is performed by comparing the correlation between the number of times the ends of the free DNA of the sample and the number of times the ends of the free DNA of the non-cancer population appear at different sites in the marked genomic interval.

[0032] Or calculate the terminal distribution consistency by the following formula

[0033] Where n is the number of sites that appear in the terminal distribution model, a i is the frequency of site i in the terminal distribution model, b i is the frequency of site i appearing at the end of the free DNA of the sample.

[0034] In some embodiments of the present application, a method for determining the classification of a sample by means of terminal distribution consistency according to a comparison result includes:

[0035] The terminal distribution consistency (E-index) of the sample is compared with the cutoff value. When the terminal distribution consistency is greater than or equal to the cutoff value, the sample is classified as a non-cancerous sample; when the terminal distribution consistency is less than the cutoff value, the sample is classified as a cancer sample.

[0036] In some embodiments of the present application, the cancer is selected from at least one of brain cancer, head and neck cancer, lung cancer, liver cancer, breast cancer, stomach cancer, colon cancer, rectal cancer, cervical cancer, ovarian cancer, pancreatic cancer, prostate cancer, thyroid cancer, lymphoma, skin cancer, and leukemia.

[0037] In some embodiments of the present application, the method further includes obtaining at least one of the copy number variation data of the free DNA of the sample, the methylation data of the free DNA, the length distribution data of the free DNA, the terminal base distribution data of the free DNA, the marker data, and the image data for analysis, and combining the analysis results with the comparison results to determine the classification of the sample.

[0038] In a third aspect of the present application, a cancer diagnostic model is provided, which is constructed using the aforementioned method.

[0039] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the aforementioned construction method or the aforementioned classification method.

[0040] In a fifth aspect of the present application, a device is provided, which includes a processor and a memory, wherein the memory stores a computer program that can be executed on the processor, and the processor implements the aforementioned construction method or the aforementioned classification method when executing the computer program.

[0041] In a sixth aspect of the present application, a device for constructing a cancer diagnosis model is provided, the device comprising:

[0042] An acquisition module, which is used to acquire sequencing data of free DNA from non-cancer patients;

[0043] A comparison module is used to compare the cell-free DNA sequencing data of non-cancer patients with the reference genome to obtain the corresponding sites of the ends of the cell-free DNA on the reference genome;

[0044] The model building module is used to build a cancer diagnosis model based on the corresponding sites of the ends of free DNA of non-cancer people on the target genome interval.

[0045] In a seventh aspect of the present application, a device for diagnosing cancer is provided, the device for diagnosing cancer comprising:

[0046] An acquisition module, which is used to acquire terminal distribution data of free DNA of a sample;

[0047] The comparison module is used to compare the free DNA end distribution data of the sample with the free DNA end distribution data of the non-cancer population, and determine whether the sample is a cancer sample based on the comparison result.

[0048] From the above, it can be seen that this application proposes a new method for cancer diagnosis based on the terminal distribution pattern of peripheral blood cfDNA. The experiments required to obtain cfDNA data can use the commonly used whole genome sequencing (whole genome shotgun sequencing), which is simple to operate and does not require high initial materials; and the diagnostic model can utilize most of the data information, does not require high sequencing depth, and has low diagnostic costs. In addition, this technology has the following characteristics:

[0049] 1) Not using cancer patient samples to identify tumor markers;

[0050] 2) Failure to use statistical methods to identify tumor markers;

[0051] 3) The diagnostic model can be used in multiple cancer types (i.e., pan-cancer cancer diagnosis can be achieved);

[0052] 4) The diagnostic model has no specific requirements for the chemical reagents and sequencing platforms used in the experiment (strong versatility).

[0053] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of a method for constructing a cancer diagnosis model provided in an embodiment of the present application.

[0055] Figure 2 It is a flowchart of a sample classification method provided in an embodiment of the present application.

[0056] Figure 3 It is a schematic diagram of the entire process of cancer diagnosis in Example 1 of the present application.

[0057] Figure 4 It is the comparison result of the E-index calculated for the samples of data set 1 in Example 1 of the present application.

[0058] Figure 5 It is the ROC curve of the E-index of the liver cancer / lung cancer in Dataset 1 and the control group in Example 1 of the present application.

[0059] Figure 6 This is the correlation test result between the E-index and tumor DNA abundance of the cancer patients in Dataset 1 in Example 1 of the present application.

[0060] Figure 7 It is the comparison result of the E-index calculated for the samples of data set 2 in Example 1 of the present application.

[0061] Figure 8 1 is the ROC curve of the E-index of different cancers and the control group in Dataset 2 in Example 1 of the present application.

[0062] Fig. 9 It is the comparison result of the E-index of lung cancer with and without metastasis in data set 2 in Example 1 of the present application.

[0063] Fig.10 It is the ROC curve of the E-index of lung cancer with and without metastasis in data set 2 in Example 1 of the present application.

[0064] Fig.11 It is the comparison result of the E-index calculated for the samples of data set 3 in Example 1 of the present application.

[0065] Fig.12 It is the ROC curve of the E-index of different cancers and the control group in Dataset 3 in Example 1 of the present application. DETAILED DESCRIPTION

[0066] The following will clearly and completely describe the concept of the present application and the technical effects produced in combination with the embodiments, so as to fully understand the purpose, features and effects of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of them. Based on the embodiments of the present application, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present application.

[0067] The embodiments of the present application are described in detail below. The described embodiments are exemplary and are only used to explain the present application, and should not be understood as limiting the present application.

[0068] In the description of this application, "several" means more than one, "many" means more than two, "greater than", "less than", "exceed", etc. are understood to exclude the number itself, and "above", "below", "within", etc. are understood to include the number itself. If there is a description of the first and the second, it is only for the purpose of distinguishing the technical features, and it cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features. If a logical order is described in the flowchart, in some cases, the steps described or shown can be performed in a different order from that in the flowchart.

[0069] In the description of the present application, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0070] This application provides a method for constructing a cancer diagnosis model, referring to Figure 1 , the method comprises the following steps:

[0071] S110, obtain sequencing data of free DNA of non-cancer population. Specifically, non-cancer population refers to a collection of people who do not have cancer, for example, people who have been confirmed to have never had cancer by one or more methods such as tissue, cell, molecule, immunity, imaging, etc. Specifically, non-cancer population can be healthy people or people with other diseases other than cancer (such as basic diseases, including hypertension, diabetes, etc.). There is no restriction on the number of non-cancer population, for example, it can be 1 to 10,000 or more people, further can be more than 5 people, more than 10 people, more than 20 people, more than 30 people, more than 50 people, more than 100 people, specifically can be 1, 2, 5, 10, 20, 50, 100, 200, 500, 1000, 2000, 5000, 10000 people. Free DNA (or cfDNA) refers to plasma free DNA, usually refers to nucleic acid fragments outside cells in peripheral blood circulation.

[0072] S120, the sequencing data of free DNA is compared with the reference genome, and the site on the reference genome corresponding to the end of the free DNA is obtained. In this step, the free DNA is positioned on the reference genome by comparing the sequencing data of the free DNA with the reference genome, and the positioning site of the end of the free DNA on the reference genome is obtained. Specifically, the reference genome is an optional human reference genome, including but not limited to any one of GRCh36 (hg18), GRCh37 (hg19), GRCh38 (hg38), etc. Of course, the reference genome is not limited to this, and can also be a specific interval on the human reference genome, and only a part of the interval of interest is compared when comparing. In addition, the genome of one or more human individuals or a specific interval thereon can also be considered. It is understandable that, depending on the difference in sequencing methods, the end of the free DNA can refer to one end or both ends of the free DNA. Specifically, when single-end sequencing is used, only the positioning site of a single end on the reference genome is selected in this step; while in double-end sequencing, the end of the free DNA can select any one of the upstream end, the downstream end, or both ends on the reference genome to participate in the construction of the model. The upstream end and the downstream end refer to the upstream or downstream of cfDNA sequencing. The end and the positioning site usually refer to the base at the end and the site corresponding to the base on the reference genome.

[0073] S130, construct a cancer diagnosis model according to the corresponding sites of the ends of the free DNA in the target genome interval. Specifically, the cancer diagnosis model includes a terminal distribution model of free DNA in non-cancer populations, and the terminal distribution model of free DNA in non-cancer populations refers to a model that characterizes the terminal distribution pattern of free DNA in non-cancer populations, and this terminal distribution pattern refers to the distribution characteristics of the ends of free DNA in non-cancer populations at different sites in the target genome interval. In some specific embodiments, this distribution feature is reflected by the frequency of occurrence of the ends of free DNA at different sites in the target genome interval. For this reason, according to the corresponding sites of the ends of free DNA in the target genome interval, the method for constructing a terminal distribution model of free DNA in non-cancer populations includes: calculating the number of times different sites in the target genome interval appear as the ends of free DNA, constructing a terminal distribution model that characterizes the site-frequency relationship, and using this terminal distribution model of cfDNA in non-cancer populations as a cancer diagnosis model. In addition, the target genome interval can be the whole human genome, or a set interval in the whole human genome, such as a partial chromosome, for example, any one or more of chr1-22, chrX, chrY, or one or more custom intervals in the whole human genome. It should be noted that if a set interval in the whole human genome is used, it can be achieved by at least one or more of the following methods, such as using capture or enrichment technology (such as target capture, whole exome sequencing, etc.) during sequencing before acquiring data, or only selecting DNA sequencing reads (sequencing reads) of the set interval when acquiring data. The reference genome and the target genome interval can be the same or different, but obviously the two need to have overlapping parts to complete the construction of the model, and in some cases, the reference genome covers the target genome interval.

[0074] In the peripheral blood of non-cancer populations, cfDNA usually comes from normal cells, and the overall fragmentation pattern is relatively similar and highly consistent. In cancer patients, due to the high heterogeneity of the tumor itself, the fragmentation patterns of cfDNA released by tumors of different cancer types, or even different patients of the same cancer type, may be very different, resulting in the tumor markers identified based on tumor patient samples generally having the dilemma of low universality and low sensitivity. In the embodiments of the present application, the applicant takes another perspective. Although the cfDNA fragmentation patterns of each tumor patient may be different from each other, there is a commonality, that is, they are all different from the DNA fragmentation patterns of non-cancer populations. Based on this principle, an idea opposite to the mainstream method is adopted. Instead of trying to identify diagnostic markers based on tumor cfDNA-specific fragmentation patterns, the terminal distribution pattern of cfDNA in non-cancer populations is determined, and the terminal distribution model of free DNA in non-cancer populations is used as a cancer diagnosis model. By comparing the sample cfDNA with the model, if the difference is large, the sample can be classified as a cancer sample, thereby distinguishing patients with various different cancers from the control population. It has good specificity and sensitivity, and achieves pan-cancer diagnosis.

[0075] The present application also provides a cancer diagnosis model constructed according to the above method. Unlike common tumor markers, this model describes the terminal distribution pattern of cfDNA released by non-tumor cells (or "normal" or "background") in peripheral blood. The fragmentation patterns of these non-tumor-derived cfDNA are relatively similar, so the models established using samples from different types of people who do not have cancer are highly consistent and are not affected by tumor heterogeneity. Compared with the identification of tumor-specific markers, it has higher universality and provides strong support for pan-cancer diagnostic technology.

[0076] In addition, the model does not need to screen biomarkers through statistical analysis, but instead incorporates all data into the model. This is because when the amount of data is large enough, the biological laws behind it have made it better conform to specific distribution laws, and there is no need for statistical analysis for screening. This method can reduce the complexity of model building, avoid various errors caused by improper use of statistical models, and at the same time improve data utilization and reduce the requirements for sequencing depth.

[0077] This application also provides a sample classification method, refer to Figure 2 , the classification method includes the following steps:

[0078] S210, obtain the free DNA end distribution data of the sample. The free DNA end distribution data refers to the positioning of the end of the free DNA in the target genome interval, specifically, it refers to the distribution pattern of the positioning sites of one or two ends of the multiple free DNAs in the sample in the target genome interval described in mathematical language. To this end, the method for obtaining the free DNA end distribution data of the sample includes: obtaining the sequencing data of the free DNA of the sample; comparing the sequencing data of the free DNA of the sample with the reference genome to obtain the corresponding sites of the ends of each free DNA in the sample on the reference genome; counting the corresponding sites of the ends of different free DNAs on the target genome interval, and calculating the number of times different sites in the target genome interval appear as the ends of the free DNA of the sample. In this step, the reference genome or the target genome interval can be an optional human reference genome, including but not limited to any one of GRCh36 (hg18), GRCh37 (hg19), GRCh38 (hg38), etc. Of course, the two are not limited to this, and can also be a specific interval on the human reference genome or the genome of a human individual or a specific interval thereon, and its specific examples include but are not limited to partial chromosomes, such as any one or more of chr1-22, chrX, chrY, or one or more customized intervals in the human whole genome. The reference genome and the target genome interval can be the same or different, but obviously the two need to have overlapping parts, and in some cases, the reference genome covers the target genome interval. At the same time, depending on the sequencing method, the end of the free DNA can refer to one end or both ends of the free DNA.

[0079] S220, compare the free DNA end distribution data of the sample with the free DNA end distribution data of the non-cancer population. In this step, the free DNA end distribution data of the non-cancer population is obtained by the following method: obtaining the sequencing data of the free DNA of the non-cancer population; comparing the sequencing data of the free DNA of the non-cancer population with the reference genome to obtain the corresponding sites of the ends of the free DNA of the non-cancer population on the reference genome; calculating the number of times different sites in the target genome interval appear as the ends of the free DNA of the non-cancer population, obtaining a statistical table of the number of times each site appears as the end of the free DNA, and constructing the terminal distribution model that characterizes the site-frequency relationship. For details, please refer to the first aspect of the embodiment of the present application.

[0080] In some specific embodiments, the method for comparing the free DNA end distribution data of the sample and the non-cancer population includes comparing the consistency of the number of times different sites in the target genomic interval appear as the ends of the free DNA of the sample and the number of times they appear as the ends of the free DNA of the non-cancer population. Specifically, the consistency comparison can be calculated by the formula Where n is the number of sites that appear in the terminal distribution model, a i is the frequency of site i in the terminal distribution model, b i is the frequency of site i appearing in the free DNA end of the sample. In this implementation, the free DNA location site and the frequency of occurrence as the end are directly used as the end distribution model, and the model of the non-cancer population is used as the weight to weight the frequency of occurrence of the corresponding site in the sample to count the number of occurrences of all sites as the index E-index reflecting the consistency of the end distribution in the sample and the model. It can be understood that since the DNA end sites that only appear in the sample or the non-cancer population must have a i or b i In the case where one of the two is 0, n can also only represent the number of free DNA end sites in the sample that overlap with the free DNA end sites in the free DNA end distribution model of the non-cancer population, thereby only calculating the n sites that overlap between the two. In addition, the above weighted results can be further normalized. For example, the total number of free DNA in the sample (N) can be used to calculate the number of free DNA end sites in the sample. 样本 ) The above results are normalized once as E-index, that is, Furthermore, the total number of free DNA in the terminal distribution model (N 模型 ) is normalized twice as E-index, It can be understood that in this step, in addition to the weighted average model, a suitable statistical method can also be used to compare the free DNA end site-frequency of the sample in the target genome interval with the free DNA end site-frequency of the non-cancer population to determine the consistency between the two. For example, the correlation of the frequency data of these two groups of specific sites is compared by a suitable non-parametric test method, such as Pearson correlation, Spearman correlation, etc., to determine the consistency of the terminal distribution of the sample and the model.

[0081] It should be noted that the terminal distribution consistency E-index value is a quantitative method of the cfDNA fragmentation model. It can be used alone for cancer diagnosis or in combination with other biomarkers (including quantitative methods of different cfDNA fragmentation patterns, as well as markers such as copy number variation that are not related to the cfDNA fragmentation pattern).

[0082] S230, determine the classification of the sample according to the comparison result. Specifically, the method for determining the classification of the sample according to the comparison result includes: comparing the E of the sample with the cutoff value, when the E-index is greater than or equal to the cutoff value, the sample is classified as a non-cancerous sample; when the E-index is less than the cutoff value, the sample is classified as a cancer sample. Among them, the cutoff value (cut off value) is the boundary value for judging the positive and negative of the test to distinguish between normal and abnormal. In the embodiment of the present application, different samples are distinguished as belonging to a control population without cancer or a population with cancer. The determination of the cutoff value can be carried out by any method well known in the art, among which the most commonly used is the receiver operating characteristic curve (ROC curve).

[0083] It is understandable that in the sample classification method provided in the embodiment of the present application, the discrimination model used is constructed by a non-cancer population, and thus, a variety of different cancers can be distinguished therefrom. The specific type of cancer can be at least one selected from brain cancer, head and neck cancer, lung cancer, liver cancer, breast cancer, stomach cancer, colon cancer, rectal cancer, cervical cancer, ovarian cancer, pancreatic cancer, prostate cancer, thyroid cancer, lymphoma, skin cancer, and leukemia. Among them, brain cancer includes primary brain cancer and secondary brain cancer. Primary brain cancer includes glioma and non-glioma (such as meningioma, medulloblastoma, medullary pituitary adenoma, primary CNS lymphoma and central nervous system germ cell tumor, etc.), secondary brain cancer includes malignant tumors that originate elsewhere but have spread (metastasized) to the brain, such as breast cancer and lung cancer that have metastasized to the brain; head and neck cancer includes oral / oropharyngeal cancer, nasopharyngeal cancer, nasal / sinus cancer, laryngeal cancer / hypopharynx cancer, salivary gland cancer, etc.; lung cancer includes small cell lung cancer, non-small cell lung cancer, etc.; liver cancer includes hepatocellular carcinoma, bile duct cancer, hepatoblastoma, etc.; breast cancer includes carcinoma in situ (such as lobular carcinoma in situ, ductal carcinoma in situ), invasive breast cancer ( invasive lobular carcinoma, invasive ductal carcinoma), etc.; gastric cancer includes gastric adenocarcinoma, gastrointestinal stromal tumor, gastrointestinal leiomyosarcoma, gastrointestinal carcinoid, gastrointestinal lymphoma, etc.; cervical cancer includes squamous cell carcinoma and adenocarcinoma, etc.; ovarian cancer includes epithelial tumors, germ cell tumors, stromal tumors, etc.; pancreatic cancer includes pancreatic cancer, cystic tumors, acinar cell carcinoma, sarcoma, ampullary carcinoma, etc.; thyroid cancer includes papillary thyroid carcinoma, follicular thyroid carcinoma, Hersh cell carcinoma, etc.; lymphoma includes Hodgkin lymphoma, non-Hodgkin lymphoma, etc.; skin cancer includes non-melanoma skin cancer and melanoma skin cancer, etc.; leukemia includes acute lymphocytic leukemia, acute myeloid leukemia, chronic lymphocytic leukemia, chronic myeloid leukemia, etc.

[0084] In addition, it should be noted that in some embodiments, the classification method provided in this application can only exclude some samples as "non-cancerous", but cannot determine the type of cancer. Cancer diagnosis and tumor tracing belong to two different categories and are applicable to different objects. A pan-cancer classification method provided in the embodiments of this application is particularly suitable for people participating in routine physical examinations, and early cancer patients are screened out through cfDNA analysis. After completing the classification provided in this application, it may be necessary to further trace the tumor to determine the specific type of cancer.

[0085] To this end, the classification method may further include a step of confirming the source of the cancer to further classify the sample to complete the tracing of the cancer, or combine this classification method with other classification methods to further improve the accuracy of cancer diagnosis. Specifically, it can be analyzed in combination with other data of free DNA or analyzed in combination with other types of data of the subject. Other data of free DNA include but are not limited to copy number variation data of free DNA, methylation data of free DNA, length distribution data of free DNA, terminal base distribution data of free DNA, etc. Other types of data of the subject may be marker data, imaging data, etc., and analyzed by the model, and the classification of the sample is determined by combining the analysis results and the comparison results. The classification of the sample can be that the subject to which the sample belongs is classified into a cancer population or a non-cancer population, and further can be classified into a population with a certain specific cancer, or a population with cancer metastasis, such as lung cancer liver metastasis, etc. Among them, the marker refers to a small molecule (such as amino acids, choline, etc. as metabolic markers), nucleic acid (such as miRNA, mRNA, lncRNA, etc.), protein, flora, etc. associated with one or several specific cancer types. It can be understood that in the embodiments of the present application, the marker data is data of other types of markers that are unrelated to the copy number variation, methylation, length distribution, terminal base distribution, and terminal distribution of the above-mentioned free DNA, or other quantitative methods of free DNA as marker data.

[0086] The present application also provides a computer-readable storage medium, which stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the aforementioned construction method or the aforementioned classification method.

[0087] The present application also provides a device, which includes a processor and a memory, wherein the memory stores a computer program that can be executed on the processor, and the processor implements the aforementioned construction method or the aforementioned classification method when executing the computer program.

[0088] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs, such as the aforementioned construction method or the aforementioned classification method described in the embodiments of the present application. The processor implements the construction of a cancer diagnosis model or the classification of samples by running the non-transient software programs and instructions stored in the memory.

[0089] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; and the data storage area may store and execute the above-mentioned programs. In addition, the memory may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices.

[0090] In some embodiments of the present application, the memory may include a memory remotely disposed relative to the processor, and the remote memory may be connected to the processor via a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0091] The non-transitory software programs and instructions required to implement the above-mentioned evaluation are stored in the memory, and when executed by one or more processors, the above-mentioned cancer diagnosis model is constructed or the sample is classified.

[0092] The present application also provides a device for constructing a cancer model, wherein the device for constructing a cancer model comprises:

[0093] An acquisition module, which is used to acquire sequencing data of free DNA from non-cancer patients;

[0094] A comparison module is used to compare the cell-free DNA sequencing data of non-cancer patients with the reference genome to obtain the corresponding sites of the ends of the cell-free DNA on the reference genome;

[0095] The model building module is used to build a cancer diagnosis model based on the corresponding sites of the ends of free DNA of non-cancer people on the target genome interval.

[0096] Specifically, constructing a cancer diagnosis model based on the corresponding sites of the ends of the free DNA of non-cancer people on the target genomic interval means constructing a cancer diagnosis model based on the distribution characteristics of the ends of the free DNA of non-cancer people at different sites in the target genomic interval. For example, by calculating the number of times different sites in the target genomic interval appear as the ends of free DNA, an end distribution model that characterizes the site-frequency relationship is constructed, and this end distribution model of cfDNA of non-cancer people is used as a cancer diagnosis model.

[0097] The present application also provides a device for diagnosing cancer, the device for diagnosing cancer comprising:

[0098] An acquisition module, which is used to acquire terminal distribution data of free DNA of a sample;

[0099] The comparison module is used to compare the free DNA end distribution data of the sample with the free DNA end distribution data of the non-cancer population, and determine whether the sample is a cancer sample based on the comparison result.

[0100] Specifically, the acquisition module is used to obtain the terminal distribution data of the free DNA of the sample. The terminal distribution data refers to the number of times different sites in the target genome interval appear as the terminal of the free DNA of the sample. The terminal distribution data of the free DNA of the non-cancer population refers to the terminal distribution model constructed based on the number of times different sites in the target genome interval appear as the terminal of the free DNA of the non-cancer population. The comparison module calculates the consistency of the terminal distribution through the formula Where n is the number of sites that appear in the terminal distribution model, a i is the frequency of site i in the terminal distribution model, b i is the frequency of site i appearing in the end of the free DNA of the sample. It can be understood that in this comparison module, in addition to using the above-mentioned terminal distribution consistency E-index for comparison, a suitable statistical method can also be used to compare the results of the free DNA terminal site-frequency of the sample in the reference genome with the free DNA terminal site-frequency of the non-cancer population, so as to compare the consistency of the two.

[0101] It is understandable that the device for diagnosing cancer may also include modules related to building a cancer model, or the corresponding acquisition modules and comparison modules may have multiple functions.

[0102] Specifically, the device for diagnosing cancer includes:

[0103] An acquisition module, which is used to acquire sequencing data of free DNA of non-cancer patients and terminal distribution data of free DNA of samples;

[0104] A comparison module is used to compare the free DNA sequencing data of non-cancer population with the reference genome to obtain the corresponding sites of the free DNA ends on the reference genome; and to compare the free DNA end distribution data of the sample with the cancer diagnosis model, and determine whether the sample is a cancer sample based on the comparison result;

[0105] The model building module is used to build a cancer diagnosis model based on the corresponding sites of the ends of free DNA of non-cancer people on the target genome interval.

[0106] The device implementation described above is only illustrative, and the modules described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0107] It is understood that all or some of the steps disclosed above, etc. can be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, and the computer-readable medium can include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). It is understood that computer storage media include volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer.

[0108] Additionally, it will be appreciated that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0109] The sequencing method of the present application is described below with reference to specific examples.

[0110] Example 1

[0111] 1. Building a cancer diagnosis model

[0112] 1. Sample source

[0113] 24 volunteers were recruited from the hospital's physical examination population. The physical examination results confirmed that there were no cancer patients. All volunteers signed informed consent forms and underwent ethical review before sample collection.

[0114] 2. Collection and Processing of Peripheral Blood

[0115] Use a blood collection tube containing an anticoagulant (EDTA) to collect peripheral blood (6 ml) from the vein, centrifuge it at 4°C and 1600×g for 15 minutes, aspirate the upper liquid and transfer it to a new centrifuge tube, centrifuge it again at 4°C and 16000×g for 15 minutes, then aspirate the upper liquid (i.e., plasma), transfer it to a freezing tube, and store it in a -80°C refrigerator.

[0116] 3. Extraction and sequencing of cfDNA

[0117] Commercial kits and equipment are used to extract cfDNA from plasma, construct sequencing libraries, and perform high-throughput sequencing (double-end). The sequencing raw data is stripped of sequencing adapters and low-quality cycles, and repetitive sequences, and then aligned to the human genome, removing repetitive sequences and low-quality alignment results to obtain the final alignment results. The positions of both ends of each cfDNA on the genome are extracted, which is the terminal site information of the cfDNA. Reference Figure 3 , count the number of cfDNA ends located at each site on the reference genome, obtain a statistical table of the number of cfDNA ends for each site in the genome, and construct a site-frequency end distribution model.

[0118] 2. Sample Classification

[0119] 1. Calculation of cfDNA end distribution consistency (E-index)

[0120] For each sample to be tested, the cfDNA end information was extracted according to the method in building a cancer diagnosis model, and the total number of cfDNA fragments in the comparison results (N 样本 ).

[0121] For all sites in the genome, the number of times they are cfDNA ends in the terminal distribution model is used as the weight, and the number of times they are cfDNA ends in the sample to be tested is weighted; then N 样本 The weighted value is normalized as the E-index that characterizes the consistency between the cfDNA terminal distribution of the sample to be tested and the non-cancer population, referring to the following formula:

[0122]

[0123] Where n is the number of sites that appear in the terminal distribution model, a i is the frequency of site i in the terminal distribution model, b i is the frequency of site i at the end of the free DNA of the sample; N 样本 is the total number of cfDNA in the sample.

[0124] refer to Figure 3To further illustrate the calculation method of E-index, assume that there are 10 sites in the model, and the number of times these 10 sites appear as the ends of cfDNA in non-cancer populations is 2, 1, 1, 1, 1, 0, 1, 2, 0, 3, and assume that the number of times the 10 sites appear as the ends of cfDNA in the sample is 1, 0, 0, 0, 1, 1, 0, 0, 1, 2, and the frequency in the model is multiplied by the frequency in the sample as the weight and normalized by the number of cfDNA to be tested in the sample (3). The calculation process is as follows:

[0125] E-index=(2×1+1×0+1×0+1×0+1×1+0×1+1×0+2×0+0×1+3×2) / 3=3.

[0126] 2. Determination of cutoff value

[0127] refer to Figure 3 , recruit a group of non-cancer patients (not overlapping with the patients recruited when building the model) as the control group and a group of cancer patients (can be a single cancer type or multiple cancer types) as the cancer group, and calculate the E-index value of cfDNA in each patient sample. Due to the presence of cfDNA from tumors in cancer patients, the E-index value of cancer patients should be lower than that of non-cancer patients overall, so a cutoff value can be used to distinguish them, thereby achieving cancer diagnosis.

[0128] The receiver operating characteristic curve (ROC) is used to evaluate the accuracy, sensitivity, specificity and other indicators when using different cutoff values ​​as diagnostic parameters for cancer diagnosis, and the cutoff value is determined according to actual needs. For example, community screening may require specificity > 99%, while high-risk groups may require sensitivity > 95%. In this way, the cutoff value required to meet specific needs can be determined.

[0129] Use the AUC (Area Under the Curve) value to evaluate the overall effect of classification: the higher the AUC, the better the effect.

[0130] 3. Classification of samples

[0131] After determining the cutoff value, the E-index value of the cfDNA is calculated. If it is lower than the cutoff value, the patient is classified as a cancer patient, otherwise he is classified as a non-cancer patient.

[0132] The samples were classified using data sets from different sources, and the results are as follows:

[0133] (1) Dataset 1 contains sample data of 10 people without cancer, 10 patients with liver cancer, and 10 patients with lung cancer (dataset source reference Liang et al. Whole-genome sequencing of cell-free DNA yields genome-wide read distribution patterns to track tissue of origin in cancer patients. Clin Transl Med. 2020 Oct; 10(6): e177.).

[0134] According to the consistency calculation method provided above, the E-index values ​​of these people are calculated as follows: Figure 4 As shown in the figure, it can be seen that compared with the people without cancer as the control group, the E-index values ​​of liver cancer patients were significantly lower overall (P<0.05), and the same was true for lung cancer patients (P<0.01).

[0135] ROC analysis results reference Figure 5 , the upper part is the ROC curve of lung cancer, and the lower part is the ROC curve of liver cancer. When using E-index for cancer diagnosis, the AUC value in liver cancer patients reaches 0.77, and the AUC value in lung cancer patients reaches 0.91, indicating that E-index can effectively distinguish cancer patients from those who do not have cancer. When the specificity requirement is 0.90, the sensitivity of cancer diagnosis is 0.70 (liver cancer) and 0.80 (liver cancer).

[0136] The ichorCNA algorithm was used to detect cfDNA somatic copy number variations (SCNA) in the samples of the above 20 patients and quantitatively calculate tumor DNA abundance (TFx), and to test the correlation between tumor DNA abundance and consistent E-index. The results are as follows Figure 6 As shown in the figure, it can be seen that in cancer patients, the E-index value is negatively correlated with the proportion of cfDNA from tumors. Therefore, the E-index can be used to evaluate the size, malignancy and other information of the patient's tumor.

[0137] (2) Dataset 2 contains sample data of 7 people without cancer (some of whom are hepatitis B carriers), 4 brain tumor patients, 4 colorectal cancer patients, 4 gastric cancer patients, 3 liver cancer patients, 15 lung cancer patients, and 7 pancreatic cancer patients (dataset source reference Song et al. 5-Hydroxymethylcytosine signatures in cell-free DNA provide information about tumor types and stages. Cell Res. 2017 Oct; 27(10): 1231-1242.).

[0138] After calculating the E-index values ​​of these people according to the consistency calculation method provided above, due to the small sample size of some cancer patients, they were first integrated to compare the cancer group (brain tumor + colon cancer + stomach cancer + liver cancer + lung cancer + pancreatic cancer) and the control group (people without cancer). The overall E-index values ​​of the two groups were significantly different (Mann-Whitney Rank-sumtest, P=0.0064).

[0139] Then, the patients were grouped according to cancer type and the E-index of different cancer types was compared with that of the control group. The results are as follows: Figure 7 As shown, the E-index of patients with most cancer types is lower than that of people without cancer.

[0140] We selected several cancer types with a large number of patients, including gastrointestinal cancer patients (including colorectal cancer patients and gastric cancer patients), pancreatic cancer patients and lung cancer patients, and performed ROC analysis on them. Figure 8 As shown in the figure, in the specificity range of 0.8 to 0.6, from top to bottom are the ROC curves of lung cancer, gastrointestinal cancer and pancreatic cancer. It can be seen from the figure that the AUCs of cancer diagnosis using E-index are 0.80 (gastrointestinal cancer), 0.78 (pancreatic cancer) and 0.88 (lung cancer), respectively, indicating that E-index can effectively distinguish cancer patients from those who do not have cancer. When the specificity requirement is 0.85, the sensitivity is 0.75 (gastrointestinal cancer), 0.71 (pancreatic cancer) and 0.80 (lung cancer), respectively.

[0141] In addition, among the 15 lung cancer patients, some had already developed tumor metastasis (liver metastasis). Fig. 9 The E-index value of these patients was also significantly lower than that of lung cancer patients without metastasis. Fig.10The AUC value of E-index for patients with liver metastasis and those without metastasis was 0.91. These results indicate that E-index can be used to predict whether a tumor will metastasize, especially to predict lung cancer metastasis, especially to predict liver metastasis of lung cancer.

[0142] (3) Dataset 3 contains sample data of 231 people without cancer, 26 patients with bile duct cancer, 54 patients with breast cancer, 27 patients with colorectal cancer, 27 patients with gastric cancer, 79 patients with lung cancer, and 28 patients with ovarian cancer (dataset source reference Cristiano et al. Genome-wide cell-free DNA fragmentation in patients with cancer. Nature 2019; 570: (385–389).).

[0143] The E-index values ​​of these people were calculated according to the consistency calculation method provided above, and the differences in the E-index values ​​of different types were compared. The results are as follows: Fig.11 As shown in the figure, it can be seen that the E-index of these cancer patients is lower than that of non-cancer patients. The ROC analysis results are as follows Fig.12 As shown in the figure, when the specificity is 0.6, from top to bottom are the ROC curves of bile duct cancer, gastric cancer, intestinal cancer, breast cancer, lung cancer, and ovarian cancer. It can be seen from the figure that when using E-index for cancer diagnosis, the AUC value is between 0.62 and 0.83, indicating that E-index can effectively distinguish cancer patients from those who do not have cancer. When the specificity requirement is 0.90, the sensitivity is between 0.15 and 0.7.

[0144] Example 2

[0145] This embodiment provides a device for diagnosing cancer, the device comprising:

[0146] An acquisition module is used to obtain a site-frequency statistical table consisting of the number of times the ends of the free DNA of the sample appear at each site in the human genome as its end distribution data;

[0147] The comparison module compares the site-frequency of the sample with the site-frequency of the diagnostic model, tests the consistency of the two through the Pearson correlation coefficient, and determines whether the subject has cancer based on the cutoff value of R in the model.

[0148] Among them, the site-frequency of the diagnostic model was constructed by comparing the sequencing results of free DNA samples from 1,000 non-cancer people.

[0149] The device for diagnosing cancer can achieve similar effects of diagnosing cancer in different data sets as in Example 1, and will not be described in detail here.

[0150] Example 3

[0151] This embodiment provides a device for diagnosing cancer, the device comprising:

[0152] An acquisition module is used to obtain a site-frequency statistical table consisting of the number of times the ends of the free DNA of the sample appear at each site in the human genome as its end distribution data; and to obtain the copy numbers of multiple genes of the free DNA of the sample as copy number variation data;

[0153] The comparison module compares the site-frequency of the sample with the site-frequency of the diagnostic model, calculates the E-index value, calculates the CNV risk score based on the copy number variation data, and combines the E-index value and the CNV risk score to determine whether the subject has cancer. Among them, the selection of genes and the calculation method of the CNV risk score can be obtained from the literature and will not be repeated here.

[0154] It can be seen from the above embodiments that the cancer diagnosis model and the corresponding classification method provided in this application can be used for a variety of cancer types, with high universality and a wide range of application scenarios. At the same time, the diagnostic accuracy of the above method is relatively high, especially in early cancer patients compared with traditional methods. In addition, the data used to implement the above scheme only need to be obtained by commonly used whole genome sequencing, and there are no special requirements for experimental schemes and sequencing platforms. The operation complexity is low and the cost is low, which is suitable for clinical promotion.

[0155] The present application is described in detail above in conjunction with the embodiments, but the present application is not limited to the above embodiments, and various changes can be made within the knowledge of ordinary technicians in the relevant technical field without departing from the purpose of the present application. In addition, the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

Claims

1. A device for diagnosing cancer, characterized in that The device comprises: An acquisition module, the acquisition module is used to: acquire sequencing data of free DNA of the sample; compare the sequencing data of the free DNA of the sample with the reference genome to obtain the corresponding sites of the ends of the free DNA of the sample on the reference genome; calculate the number of occurrences of different sites in the target genome interval as the corresponding sites of the ends of the free DNA of the sample; A comparison module, wherein the comparison module is used to compare the free DNA end distribution data of the sample with the free DNA end distribution data of a non-cancer population, and determine whether the sample is a cancer sample based on the comparison result; the comparison includes comparing the consistency of the number of times different sites in the target genome interval appear as the ends of the free DNA of the sample and the number of times they appear as the ends of the free DNA of the non-cancer population.

2. The device according to claim 1, characterized in that The free DNA end distribution data of the non-cancer population is obtained by the following method: Obtain sequencing data of cell-free DNA from non-cancer populations; The sequencing data of the free DNA of the non-cancer population is compared with the reference genome to obtain the corresponding sites of the ends of the free DNA of the non-cancer population on the reference genome; the number of occurrences of different sites in the target genome interval as the corresponding sites at the ends of the free DNA of the non-cancer population is calculated, and an end distribution model that characterizes the site-frequency relationship is constructed.

3. The device according to claim 2, characterized in that The consistency is compared by comparing the correlation of the two times or calculating the terminal distribution consistency by the following formula: Where n is the number of sites appearing in the terminal distribution model, a i is the frequency of occurrence of site i in the terminal distribution model, b i is the frequency of occurrence of site i at the end of the free DNA of the sample.

4. The device according to claim 3, characterized in that The correlation comparison includes at least one of Pearson correlation and Spearman correlation.

5. The device according to claim 3, characterized in that The calculation of the terminal distribution consistency is normalized by the total number of free DNA in the sample as the E-index, that is, Among them, N 样本 is the total number of free DNA in the sample.

6. The device according to claim 5, characterized in that The calculation of the terminal distribution consistency is performed by normalizing the total number of free DNA in the terminal distribution model twice as the E-index, that is, Among them, N 模型 is the total number of free DNA in the terminal distribution model.

7. The device according to claim 1, characterized in that The method for determining whether the sample is a cancer sample according to the comparison result comprises: The terminal distribution consistency of the sample is compared with the cutoff value. When the terminal distribution consistency is greater than or equal to the cutoff value, the sample is classified as a non-cancerous sample; when the terminal distribution consistency is less than the cutoff value, the sample is classified as a cancer sample.

8. The device according to claim 1, characterized in that The cancer is selected from at least one of brain cancer, head and neck cancer, lung cancer, liver cancer, breast cancer, stomach cancer, colon cancer, rectal cancer, cervical cancer, ovarian cancer, pancreatic cancer, prostate cancer, thyroid cancer, lymphoma, skin cancer, and leukemia.

9. The device according to claim 1, characterized in that The end of the free DNA includes at least one of an upstream end and a downstream end.

10. The device according to claim 1, characterized in that The target genome interval is the human whole genome or a set interval of the human whole genome.

11. The device according to any one of claims 1 to 10, characterized in that The comparison module also includes obtaining at least one of the copy number variation data of the free DNA of the sample, the methylation data of the free DNA, the length distribution data of the free DNA, the terminal base distribution data of the free DNA, the marker data, and the image data for analysis, and combining the analysis result with the comparison result to determine whether the sample is a cancer sample.

12. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute a classification method, and the classification method includes: Acquire sequencing data of the free DNA of the sample; compare the sequencing data of the free DNA of the sample with the reference genome to obtain corresponding sites of the ends of the free DNA of the sample on the reference genome; calculate the number of occurrences of different sites in the target genome interval as the corresponding sites of the ends of the free DNA of the sample; The free DNA end distribution data of the sample is compared with the free DNA end distribution data of a non-cancer population, and whether the sample is a cancer sample is determined based on the comparison result; the comparison includes comparing the consistency of the number of times different sites in the target genomic interval appear as the ends of the free DNA of the sample and the number of times they appear as the ends of the free DNA of the non-cancer population.

13. The computer-readable storage medium according to claim 12, wherein: The free DNA end distribution data of the non-cancer population is obtained by the following method: Obtain sequencing data of cell-free DNA from non-cancer populations; The sequencing data of the free DNA of the non-cancer population is compared with the reference genome to obtain the corresponding sites of the ends of the free DNA of the non-cancer population on the reference genome; the number of occurrences of different sites in the target genome interval as the corresponding sites at the ends of the free DNA of the non-cancer population is calculated, and an end distribution model that characterizes the site-frequency relationship is constructed.

14. The computer-readable storage medium according to claim 13, wherein: The consistency is compared by comparing the correlation of the two times or calculating the terminal distribution consistency by the following formula: Where n is the number of sites appearing in the terminal distribution model, a i is the frequency of occurrence of site i in the terminal distribution model, b i is the frequency of occurrence of site i at the end of the free DNA of the sample.

15. The computer-readable storage medium according to claim 14, wherein: The correlation comparison includes at least one of Pearson correlation and Spearman correlation.

16. The computer-readable storage medium according to claim 14, wherein: The calculation of the terminal distribution consistency is normalized by the total number of free DNA in the sample as the E-index, that is, Among them, N 样本 is the total number of free DNA in the sample.

17. The computer-readable storage medium according to claim 16, wherein: The calculation of the terminal distribution consistency is performed by normalizing the total number of free DNA in the terminal distribution model twice as the E-index, that is, Among them, N 模型 is the total number of free DNA in the terminal distribution model.

18. The computer-readable storage medium according to claim 12, wherein: The method for determining whether the sample is a cancer sample according to the comparison result comprises: The terminal distribution consistency of the sample is compared with the cutoff value. When the terminal distribution consistency is greater than or equal to the cutoff value, the sample is classified as a non-cancerous sample; when the terminal distribution consistency is less than the cutoff value, the sample is classified as a cancer sample.

19. The computer-readable storage medium according to claim 12, wherein: The cancer is selected from at least one of brain cancer, head and neck cancer, lung cancer, liver cancer, breast cancer, stomach cancer, colon cancer, rectal cancer, cervical cancer, ovarian cancer, pancreatic cancer, prostate cancer, thyroid cancer, lymphoma, skin cancer, and leukemia.

20. The computer-readable storage medium according to claim 12, wherein: The end of the free DNA includes at least one of an upstream end and a downstream end.

21. The computer-readable storage medium according to claim 12, wherein: The target genome interval is the human whole genome or a set interval of the human whole genome.

22. The computer-readable storage medium according to any one of claims 12 to 21, characterized in that: The comparison also includes obtaining at least one of the copy number variation data of the free DNA of the sample, the methylation data of the free DNA, the length distribution data of the free DNA, the terminal base distribution data of the free DNA, the marker data, and the image data for analysis, and combining the analysis result with the comparison result to determine whether the sample is a cancer sample.

23. The device, characterized in that The invention comprises a processor and a memory, wherein the memory stores a computer program that can be run on the processor, and the processor implements a classification method when running the computer program, wherein the classification method comprises: Acquire sequencing data of the free DNA of the sample; compare the sequencing data of the free DNA of the sample with the reference genome to obtain corresponding sites of the ends of the free DNA of the sample on the reference genome; calculate the number of occurrences of different sites in the target genome interval as the corresponding sites of the ends of the free DNA of the sample; The free DNA end distribution data of the sample is compared with the free DNA end distribution data of a non-cancer population, and whether the sample is a cancer sample is determined based on the comparison result; the comparison includes comparing the consistency of the number of times different sites in the target genomic interval appear as the ends of the free DNA of the sample and the number of times they appear as the ends of the free DNA of the non-cancer population.

24. The device according to claim 23, characterized in that The free DNA end distribution data of the non-cancer population is obtained by the following method: Obtain sequencing data of cell-free DNA from non-cancer populations; The sequencing data of the free DNA of the non-cancer population is compared with the reference genome to obtain the corresponding sites of the ends of the free DNA of the non-cancer population on the reference genome; the number of occurrences of different sites in the target genome interval as the corresponding sites at the ends of the free DNA of the non-cancer population is calculated, and an end distribution model that characterizes the site-frequency relationship is constructed.

25. The device according to claim 24, characterized in that The consistency is compared by comparing the correlation of the two times or calculating the terminal distribution consistency by the following formula: Where n is the number of sites appearing in the terminal distribution model, a i is the frequency of occurrence of site i in the terminal distribution model, b i is the frequency of occurrence of site i at the end of the free DNA of the sample.

26. The device according to claim 25, characterized in that The correlation comparison includes at least one of Pearson correlation and Spearman correlation.

27. The device according to claim 25, characterized in that The calculation of the terminal distribution consistency is normalized by the total number of free DNA in the sample as the E-index, that is, Among them, N 样本 is the total number of free DNA in the sample.

28. The device according to claim 27, characterized in that The calculation of the terminal distribution consistency is performed by normalizing the total number of free DNA in the terminal distribution model twice as the E-index, that is, Among them, N 模型 is the total number of free DNA in the terminal distribution model.

29. The device according to claim 23, characterized in that The method for determining whether the sample is a cancer sample according to the comparison result comprises: The terminal distribution consistency of the sample is compared with the cutoff value. When the terminal distribution consistency is greater than or equal to the cutoff value, the sample is classified as a non-cancerous sample; when the terminal distribution consistency is less than the cutoff value, the sample is classified as a cancer sample.

30. The device according to claim 23, characterized in that The cancer is selected from at least one of brain cancer, head and neck cancer, lung cancer, liver cancer, breast cancer, stomach cancer, colon cancer, rectal cancer, cervical cancer, ovarian cancer, pancreatic cancer, prostate cancer, thyroid cancer, lymphoma, skin cancer, and leukemia.

31. The device according to claim 23, characterized in that The end of the free DNA includes at least one of an upstream end and a downstream end.

32. The device according to claim 23, characterized in that The target genome interval is the human whole genome or a set interval of the human whole genome.

33. The device according to any one of claims 23 to 32, characterized in that The comparison also includes obtaining at least one of the copy number variation data of the free DNA of the sample, the methylation data of the free DNA, the length distribution data of the free DNA, the terminal base distribution data of the free DNA, the marker data, and the image data for analysis, and combining the analysis result with the comparison result to determine whether the sample is a cancer sample.

Images