Tissue tracing method based on cfDNA methylation sequencing, tissue tracing device and electronic equipment

Abstract

The invention discloses a tissue tracing method based on cfDNA sequencing. The method comprises the following steps: collecting cfDNA samples of a plurality of cancer patients and healthy people, and carrying out methylation sequencing to obtain a plurality of methylation intervals; calculating a saliency value of each methylation interval in the plurality of methylation intervals, deleting the corresponding methylation interval in response to the saliency value not lower than a first threshold value, reserving the corresponding methylation interval in response to the saliency value lower than the first threshold value, and screening the plurality of methylation intervals to obtain a candidate interval; extracting whole genome fragment information based on the reference genome, uniformly dividing the whole genome fragment information into a plurality of fragment intervals, and calculating fragment coverage of each fragment interval in the plurality of fragment intervals; a methylation model is constructed through a plurality of candidate intervals, a fragmentation model is constructed through fragment coverage, and the methylation model and the fragmentation model are integrated to obtain a joint traceability model which is used for cancer tissue traceability of various cancer patients. According to the method, the sequencing cost and the data processing complexity are reduced, and a high-accuracy tumor tissue tracing scheme is provided.

Description

Technical Field

[0001] The present application belongs to the field of molecular biomedical technology, and specifically, relates to a tissue tracing method, a tissue tracing device and an electronic device based on cfDNA methylation sequencing. Background Art

[0002] Cancer is one of the leading causes of death among residents worldwide and a major public health issue that seriously endangers human health. According to the GLOBOCAN 2020 report released by the World Health Organization's International Agency for Research on Cancer, there were 19.3 million new cases of cancer and 9.96 million cancer deaths worldwide in 2020. Studies have shown that early detection and timely diagnosis can greatly improve the survival rate of cancer patients. Taking lung cancer as an example, according to the four stages of lung cancer by the International Association for the Study of Lung Cancer, the 5-year survival rate of patients with the earliest stage (IA stage) of lung cancer can be as high as 92%, and the 5-year survival rate of patients with stage I is at least 53%; while for patients with middle and late stage lung cancer, the 5-year survival rate of patients with stage III is only 13%-36%, and the 5-year survival rate of patients with the most advanced stage (IVB stage) of lung cancer is almost zero. It can be seen that cancer screening, early diagnosis, and early treatment are of great significance to improving the survival rate of patients. In cancer detection and diagnosis, cancer tissue tracing plays a vital role. Tissue tracing technology can help determine the primary site of cancer, which is essential for formulating effective treatment strategies and improving the treatment effect of patients.

[0003] At present, a variety of tissue tracing methods have been proposed, including fragment omics feature methods based on whole genome sequencing, methylation sequencing methods based on second-generation sequencing technology, and comprehensive tracing methods combining whole genome sequencing and methylation sequencing. Although the comprehensive tracing method combining fragment omics and methylation sequencing has shown better results or potential than a single method in tracing the tissues of many cancer types, it usually requires simultaneous whole genome sequencing and high-throughput methylation sequencing, which increases the complexity and cost of tissue tracing. For example, the method SPOT-MAS developed by Nguyen, Van Thien Chi et al., which simultaneously implemented whole genome sequencing and targeted methylation sequencing and trained the model based on sequencing data, has a tracing accuracy of 0.7.

[0004] The cost of whole genome sequencing is still high, and the amount of data generated is huge, requiring more computing resources and complex bioinformatics tools for processing and analysis, which is not conducive to the practical application of tissue tracing detection. Compared with whole genome sequencing, especially whole genome methylation sequencing, targeted methylation sequencing has more advantages in tissue tracing detection limit (LoD) and cost control, which makes it more feasible and economical in clinical applications. To this end, the present application provides a method, device and electronic device that can perform tissue tracing based only on targeted methylation sequencing of cfDNA, thereby providing low-cost and good accuracy cancer tissue tracing.

[0005] Methylation features and fragmentation features were extracted simultaneously, without the need for additional whole-genome sequencing to extract fragmentation features alone. The tissue traceability accuracy was also good at 0.802. Summary of the invention

[0006] In order to solve the problems existing in the prior art, the purpose of this application is to provide a tissue tracing method, device and electronic equipment based on cfDNA methylation sequencing, so as to simultaneously extract the methylation characteristics and fragmentation characteristics in the cfDNA of the target population and synergistically act on cancer tissue tracing, thereby reducing the cost of tracing detection while maintaining high accuracy, thereby providing medical institutions and subjects with a highly practical cancer early screening and detection solution.

[0007] Specifically, this application involves the following aspects:

[0008] 1. A tissue tracing method based on cfDNA methylation sequencing, comprising: collecting cfDNA samples from various cancer patients and healthy people and obtaining multiple methylation intervals through methylation sequencing; calculating the significance value of each methylation interval in the multiple methylation intervals, deleting the corresponding methylation interval in response to the significance value being not less than a first threshold, retaining the corresponding methylation interval in response to the significance value being less than the first threshold, and screening the multiple methylation intervals to obtain multiple candidate intervals; extracting whole genome fragment information based on a reference genome and equally dividing it into multiple fragment intervals, calculating the fragment coverage of each fragment interval in the multiple fragment intervals; constructing a methylation model through multiple candidate intervals, constructing a fragmentation model through fragment coverage, integrating the methylation model and the fragmentation model to obtain a joint tracing model for cancer tissue tracing of various cancer patients.

[0009] 2. A tissue tracing method based on cfDNA methylation sequencing according to item 1, wherein screening a plurality of methylation intervals to obtain a plurality of candidate intervals comprises: respectively calculating the methylation value of each methylation interval in a plurality of cancer populations for each cancer population; respectively calculating the difference between the methylation value of each methylation interval in each cancer population and the methylation value in other cancer populations in the plurality of cancer populations, retaining the corresponding methylation interval in response to the number of methylation value differences greater than the second threshold being not less than the third threshold, and deleting the corresponding methylation interval in response to the number of methylation value differences greater than the second threshold being less than the third threshold.

[0010] 3. According to the tissue tracing method based on cfDNA methylation sequencing of item 2, wherein respectively calculating the methylation value of each methylation interval in a plurality of methylation intervals for each cancer patient population in a plurality of cancer patient populations includes: respectively calculating the mean methylation value of the CpG site in each methylation interval in a plurality of methylation intervals for each cancer patient population in a plurality of cancer patient populations.

[0011] 4. According to the tissue tracing method based on cfDNA methylation sequencing of item 1, screening multiple methylation intervals to obtain candidate intervals includes: calculating the importance value of each methylation interval in the multiple methylation intervals respectively; sorting the multiple methylation intervals according to the importance value; retaining the first n methylation intervals in the multiple methylation intervals, and deleting the remaining methylation intervals.

[0012] 5. According to the tissue tracing method based on cfDNA methylation sequencing of item 1, extracting the whole genome fragment information based on the reference genome and dividing it into multiple fragment intervals includes: tiling the whole genome fragment information according to the reference genome autosomes, and dividing it into multiple adjacent fragment intervals with an interval length of a.

[0013] 6. According to the tissue tracing method based on cfDNA methylation sequencing of item 5, calculating the fragment coverage of each fragment interval in multiple fragment intervals includes: extracting short fragments, medium fragments, long fragments and effective fragments in multiple fragment intervals, and recording the number of short fragments, medium fragments, long fragments and effective fragments; wherein, the interval length of short fragments is 100-150bp; the interval length of medium fragments is 151-260bp; the interval length of long fragments is 261-320bp; and the interval length of effective fragments is 100-320bp.

[0014] 7. According to the tissue tracing method based on cfDNA methylation sequencing of item 1, integrating the methylation model and the fragmentation model to obtain the joint tracing model includes: using the methylation model and the fragmentation model respectively to obtain the tissue tracing prediction value of the cfDNA samples of each cancer patient population in multiple cancer patient populations; based on the tissue tracing prediction value, training the joint tracing model through logistic regression.

[0015] 8. A tissue tracing device based on cfDNA methylation sequencing, comprising: a sample sequencing unit, which collects cfDNA samples from various cancer patients and healthy people and obtains multiple methylation intervals through methylation sequencing; an interval screening unit, which calculates the significance value of each methylation interval in the multiple methylation intervals, deletes the corresponding methylation interval in response to the significance value being not less than a first threshold, retains the corresponding methylation interval in response to the significance value being less than the first threshold, and screens the multiple methylation intervals to obtain multiple candidate intervals; a fragment extraction unit, which extracts whole genome fragment information based on a reference genome and divides it equally into multiple fragment intervals, and calculates the fragment coverage of each fragment interval in the multiple fragment intervals; a model construction unit, which constructs a methylation model through multiple candidate intervals, constructs a fragmentation model through fragment coverage, and integrates the methylation model and the fragmentation model to obtain a joint tracing model for cancer tissue tracing of various cancer patients.

[0016] 9. According to the tissue tracing device based on cfDNA methylation sequencing of item 8, it is a tissue tracing method based on cfDNA methylation sequencing as performed in any one of items 1-7.

[0017] 10. An electronic device, comprising: a processor; and a memory, wherein computer program instructions are stored in the memory, and when the computer program instructions are executed by the processor, the processor executes a tissue tracing method based on cfDNA methylation sequencing as described in any one of items 1-7.

[0018] 11. A tissue tracing marker, wherein the marker comprises one or more of the following: MST1: a genomic region from 49723693 to 49724001 on chromosome 3, HS3ST1: a genomic region from 11429523 to 11429690 on chromosome 4, OSMR: a genomic region from 38846033 to 38846336 on chromosome 5, PTGER4: a genomic region from 40681528 to 40681842 on chromosome 5, IRF4: a genomic region from 392623 to 392624 on chromosome 6, The genomic region at position 92926, REPIN1: the genomic region at positions 150068516 to 150068824 of chromosome 7, TMEM215: the genomic region at positions 32783168 to 32783474 of chromosome 9, PAX5: the genomic region at positions 37027033 to 37027333 of chromosome 9, KRT5: the genomic region at positions 52913892 to 52914230 of chromosome 12, and DNM2: the genomic region at positions 10823533 to 10823837 of chromosome 19.

[0019] 12. Use of a marker in preparing a kit for tissue tracing, wherein the marker comprises one or two or more of the following: MST1: a genomic region from 49723693 to 49724001 of chromosome 3, HS3ST1: a genomic region from 11429523 to 11429690 of chromosome 4, OSMR: a genomic region from 38846033 to 38846336 of chromosome 5, PTGER4: a genomic region from 40681528 to 40681842 of chromosome 5, IRF4: a genomic region from 392 The genomic region from positions 623 to 392926, REPIN1: the genomic region from positions 150068516 to 150068824 of chromosome 7, TMEM215: the genomic region from positions 32783168 to 32783474 of chromosome 9, PAX5: the genomic region from positions 37027033 to 37027333 of chromosome 9, KRT5: the genomic region from positions 52913892 to 52914230 of chromosome 12, and DNM2: the genomic region from positions 10823533 to 10823837 of chromosome 19.

[0020] 13. A composition for tissue tracing, wherein the composition comprises: a nucleic acid for detecting the methylation state of the marker of item 11; wherein the methylation state of the marker is characterized by the methylation of the target sequence of the marker.

[0021] 14. A tissue tracing kit, comprising a reagent for detecting the marker of item 11.

[0022] The tissue tracing method based on cfDNA methylation sequencing provided in the present application constructs a tissue tracing methylation model by extracting candidate intervals as methylation features in cfDNA targeted sequencing data, extracts fragmentation features based on the reference genome without implementing whole genome sequencing of sample data, and uses them to replace genome sequencing features and construct a fragmentation model. Based on the two models obtained, a joint tracing model is obtained by using an integrated method. This model has good discrimination ability and high accuracy for tissue samples of various cancers such as colorectal cancer, lung cancer, and gastric cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The figure illustrates a flow chart of a tissue tracing method based on cfDNA methylation sequencing according to an embodiment of the present application.

[0024] Figure 2 The figure illustrates a flow chart of obtaining candidate intervals according to the tissue tracing method based on cfDNA methylation sequencing according to an embodiment of the present application.

[0025] Figure 3A The figure shows a schematic diagram of the classification results of tracing the tissue origins of various cancer patients based on the methylation model of an embodiment of the present application.

[0026] Figure 3B The diagram shows a schematic diagram of the classification results of tracing the tissue origins of various cancer patients using the fragmentation model of an embodiment of the present application.

[0027] Figure 3C The diagram shows a schematic diagram of the classification results of tracing the tissues of various cancer patients using the joint tracing model according to an embodiment of the present application.

[0028] Figure 4 The figure illustrates a block diagram of a tissue tracing device based on cfDNA methylation sequencing according to an embodiment of the present application.

[0029] Figure 5 A block diagram of an electronic device according to an embodiment of the present application is illustrated. DETAILED DESCRIPTION

[0030] The present application is further described below in conjunction with examples. It should be understood that the examples are only used to further describe and illustrate the present application and are not used to limit the present application.

[0031] Unless otherwise defined, the technical and scientific terms in this specification have the same meaning as those generally understood by those skilled in the art. Although similar or identical methods and materials as described herein may be used in experiments or practical applications, materials and methods are described herein below. In the event of a conflict, the present specification including the definitions therein shall prevail. In addition, materials, methods and examples are provided for illustration only and are not restrictive. The present application is further described below in conjunction with specific embodiments, but is not intended to limit the scope of the present application.

[0032] Application Overview

[0033] As mentioned above, extracting cancer tissue markers using whole genome sequencing data requires high computational costs and application difficulties. In order to determine the primary site of the tumor, subjects often need to undergo longer examinations and pay higher fees, which affects the breadth of clinical application of tissue tracing technology. DNA methylation, as an epigenetic marker, exhibits strong tissue specificity, making it an ideal molecular marker for tracing the origin of tumor tissue.

[0034] For example, Yun Feng et al. pointed out in the paper "Comparison of the classifiers based on mRNA, microRNA and IncRNA expression and DNA methylation profiles for the tumor origin detection" that using only 1,000 highly methylated CpG sites extracted by methylation sequencing can achieve an overall accuracy of more than 90% in tissue prediction, indicating that even with low sequencing costs and limited number of markers, DNA methylation can still provide reliable tumor tissue traceability information. Using free DNA (cfDNA) in the blood for methylation analysis can achieve non-invasive early detection of cancer and tissue traceability.

[0035] Based on the above considerations, the present application provides a tissue tracing method based on cfDNA sequencing, including: collecting cfDNA samples from various cancer patients and healthy people and obtaining multiple methylation intervals through methylation sequencing; calculating the significance value of each methylation interval in the multiple methylation intervals, deleting the corresponding methylation interval in response to the significance value being not lower than a first threshold, retaining the corresponding methylation interval in response to the significance value being lower than the first threshold, and screening the multiple methylation intervals to obtain multiple candidate intervals; extracting the whole genome fragment information of the cfDNA sample based on the reference genome and dividing it equally into multiple fragment intervals, calculating the fragment coverage of each fragment interval in the multiple fragment intervals; constructing a methylation model through multiple candidate intervals, constructing a fragmentation model through fragment coverage, integrating the methylation model and the fragmentation model to obtain a joint tracing model for cancer tissue tracing of various cancer patients.

[0036] By utilizing multiple screening factors such as significance value, importance value, and expression differences of methylation levels in different cancer populations, the interval markers required for cancer tissue tracing in this application can be extracted and screened out from targeted methylation sequencing data, and these interval markers can be used to construct a cancer tissue tracing model based on gene methylation expression, and the cfDNA samples of different cancer populations and healthy populations can be classified, which can give full play to the advantages of methylation tissue tracing analysis; in addition, the reference genome is used to extract the whole genome fragment information of cfDNA samples from different populations, and the whole genome fragment information is used to capture the subtle changes in the genomic DNA fragments, which are closely related to the occurrence and development of various cancers, thereby providing rich biological information to reveal the heterogeneity between different cancer tissues, and constructing a tracing model for tissue differentiation between various cancers.

[0037] In this way, two cancer tissue tracing models with their own advantages and complementary performances are obtained. According to the tissue tracing method based on cfDNA methylation sequencing of the embodiment of the present application, the two models are combined through an optimal integration method to obtain a joint tracing model, which uses the targeted methylation sequencing data of cfDNA samples of the target population to predict the tissue origin of the tissue where the sample is located, which can significantly reduce the economic cost and time required for the detection process, provide strong support for the promotion of application products such as detection kits for corresponding markers, and promote the development of personalized tumor treatment and precision medicine.

[0038] After introducing the basic principles of the present application, various non-limiting embodiments of the present application will be described in detail with reference to the accompanying drawings.

[0039] Exemplary Methods

[0040] Figure 1 The diagram illustrates a tissue tracing method based on cfDNA methylation sequencing according to an embodiment of the present application.

[0041] like Figure 1 As shown, the tissue tracing method based on cfDNA methylation sequencing according to an embodiment of the present application includes the following steps.

[0042] Step S110, collect cfDNA samples from a variety of cancer patients and healthy people and obtain multiple methylation intervals through methylation sequencing. In the present application, the cancer patients include groups with 7 types of cancer, including lung cancer, intestinal cancer, gastric cancer, liver cancer, esophageal cancer, thyroid cancer and ovarian cancer. Collect plasma samples from the cancer patients and capture the target cfDNA samples therein, complete the establishment of methylation library and high-throughput targeted sequencing. The high-throughput targeted sequencing can be based on the sequencing products of existing public research, public testing or public use to determine the high-correlation target interval of the above-mentioned cancer, for example, using the high-throughput methylation sequencing target interval used by Borcheng's seven cancers NGS methylation detection kit and / or gastrointestinal cancer methylation joint detection kit as the target interval for sequencing.

[0043] It can be understood that in order to make the length of the target interval close to the length of human cfDNA, the length of the target interval can be limited according to the sequencing results, so that the methylation value of the CpG site in the interval is closer to the actual level, ensuring data quality; the required number of CpG sites in the target interval can also be specified to reduce the false positive probability of the methylation value of the CpG site. That is, those skilled in the art can make any beneficial limitations on the selected target interval according to actual conditions.

[0044] Furthermore, the cfDNA samples of the above-mentioned target population may include circulating tumor DNA (ctDNA) samples of the target cancer, cfDNA methylation samples, or known cfDNA and epigenome combination samples. These samples usually contain effective information such as methylation expression heterogeneity, disease status, and / or tumor size among cancer patients, and are the source of important markers for cancer diagnosis. In other words, by detecting specific mutations in the cfDNA samples of the target population, namely methylation level mutations and genomic structural variations, such as genomic fragment characteristics, etc., the cancer can be traced.

[0045] Step S120, calculate the significance value of each methylation interval in the multiple methylation intervals, delete the corresponding methylation interval in response to the significance value being not less than the first threshold, retain the corresponding methylation interval in response to the significance value being less than the first threshold, and screen the multiple methylation intervals to obtain multiple candidate intervals. Here, all methods for screening methylation intervals are based on the methylation value of each methylation interval, that is, the methylation value of the CpG site in the methylation interval, such as the average methylation value of all CpG sites in the methylation interval or the median of the methylation values ​​of all CpG sites. The screening process is to measure the difference in methylation values ​​of multiple methylation intervals. Specifically, screening multiple methylation intervals to obtain candidate intervals can also include different screening methods or a combination of these screening methods. Therefore, this step can further include the following: Figure 2 The following steps are shown.

[0046] Step S1201, calculate the significance value of each methylation interval in multiple methylation intervals, delete the corresponding methylation interval in response to the significance value being not less than the first threshold, retain the corresponding methylation interval in response to the significance value being less than the first threshold, and screen multiple methylation intervals to obtain candidate intervals. According to the tissue tracing method based on cfDNA sequencing in the embodiment of the present application, the significance value is used to determine the probability value of the possibility that each methylation interval has the same phenotype in the cancer population and the healthy population. The smaller the significance value, the more likely it is that the methylation level of the methylation interval in the cancer population and the healthy population will have different phenotypes; in this way, the methylation interval obtained based on the significance value and screened by the first threshold statistically preliminarily shows the difference between the cancer population and the healthy population.

[0047] Here, the p-value of each methylation interval in the cfDNA sample can be obtained by using the Kruskal-Wallis test, and the p-value can be corrected by using the False Discovery Rate (FDR) method to obtain the qvalue value, and the qvalue value is used as the significance value. The methylation interval with a qvalue value less than the first threshold of 0.05 is retained as the statistical difference interval between the cancer population and the healthy population. In some embodiments, the first threshold can be set to 0.05, 0.025 or 0.01. The smaller the first threshold, the more convincing the difference between the methylation intervals screened in the two populations; in a favorable embodiment, the first threshold is 0.05 to avoid interference from methylation intervals with extreme methylation values.

[0048] Step S1202, respectively calculating the methylation value of each methylation interval in each cancer patient population among the multiple cancer patient populations; calculating the difference between the methylation value of each methylation interval in each cancer patient population and the methylation value of other cancer patient populations among the multiple cancer patient populations, retaining the corresponding methylation interval in response to the number of methylation value differences greater than the second threshold being not less than the third threshold, and deleting the corresponding methylation interval in response to the number of methylation value differences greater than the second threshold being less than the third threshold. For the methylation interval, the difference in methylation values ​​between any two cancer patients in the aforementioned multiple cancer patients is calculated according to the tissue tracing method based on cfDNA sequencing in the embodiment of the present application, which is recorded as Delta. For example, the difference in methylation values ​​between the lung cancer patients and the other six cancer patients can be recorded as Lung_Colorectum_Delta, Lung_Stomach_Delta, Lung_Liver_Delta, Lung_Esophagus_Delta, Lung_Thyroid_Delta and Lung_Ovary_Delta. For each cancer patient population, the difference between each methylation interval in this cancer patient population and all other cancer patients can be calculated in this way.

[0049] In this way, when the cancer population includes k types of cancer, k*(k-1) / 2 non-repeated methylation value differences can be obtained based on each methylation interval. The larger the methylation value difference, the greater the methylation difference between the cancer populations. In order to avoid the influence of the methylation direction on the differential methylation interval, the absolute value of the methylation value difference can be used as a measurement parameter to measure the difference between the methylation intervals in different cancer populations, and it is stipulated that when the absolute value of the methylation value difference is greater than the second threshold, the methylation interval has differences between different cancer populations. Taking lung cancer as an example, for a methylation interval, after calculating its Delta with intestinal cancer, gastric cancer, liver cancer, esophageal cancer, thyroid cancer and ovarian cancer, the absolute value of Delta is taken. If the absolute value of Delta is greater than the second threshold, it is considered that the methylation interval has a difference in methylation level between lung cancer and the other cancer; in a favorable embodiment, the second threshold is 0.05.

[0050] Furthermore, if there are multiple methylation value differences greater than the second threshold value in the k-1 methylation value differences of a methylation interval, it means that the methylation interval shows different methylation levels in multiple cancers, and is a more suitable multi-classification marker for cancer tissue tracing. Therefore, the third threshold is continued to be set to screen these possible tissue tracing multi-classification markers. Taking the aforementioned seven cancer patients as an example, if a methylation interval has no less than the third threshold number of methylation value differences that meet the condition of being greater than the second threshold, the methylation interval is retained and used as a tissue tracing multi-classification marker; in a favorable embodiment, when the types of cancer patients are the aforementioned seven types, the third threshold value is 4, so that the screened methylation interval has effective characterization capabilities in at least more than half of the types of cancer tissue tracing. Perform the same operation for all methylation intervals, and finally obtain the further filtered methylation intervals of each cancer type as candidate intervals.

[0051] Step S1203, respectively calculate the importance value of each methylation interval in the multiple methylation intervals; sort the multiple methylation intervals according to the importance value; retain the first n methylation intervals in the multiple methylation intervals, and delete the remaining methylation intervals. Here, the purpose of sorting and screening the multiple methylation intervals according to the importance value is to reduce the amount of sample data of the methylation intervals, so as to reduce the fitting time of the model, and to improve the quality of the candidate intervals of the training model and reduce data redundancy. Therefore, in some possible implementation schemes, for the methylation intervals or the candidate intervals screened by step S1202, the improved random forest feature screening algorithm of the R package Boruta is used to calculate the feature importance score of each methylation interval or candidate interval as the importance value, sort the intervals according to the importance score and select the top-ranked methylation intervals or candidate intervals. In particular, this method of sorting and screening intervals by calculating the importance value can also be used for further screening of fragmented features.

[0052] In this way, multiple candidate intervals are obtained as markers for cancer tissue tracing through the method of the present application, which can be used to construct a methylation model or directly used for tissue tracing detection. In a preferred embodiment, some candidate intervals are set as tissue tracing markers, and the markers include one or more of the following:

[0053] MST1: genomic region from 49723693 to 49724001 on chromosome 3, HS3ST1: genomic region from 11429523 to 11429690 on chromosome 4, OSMR: genomic region from 38846033 to 38846336 on chromosome 5, PTGER4: genomic region from 40681528 to 40681842 on chromosome 5, IRF4: genomic region from 392623 to 392926 on chromosome 6, REPI N1: the genomic region from 150068516 to 150068824 of chromosome 7, TMEM215: the genomic region from 32783168 to 32783474 of chromosome 9, PAX5: the genomic region from 37027033 to 37027333 of chromosome 9, KRT5: the genomic region from 52913892 to 52914230 of chromosome 12, DNM2: the genomic region from 10823533 to 10823837 of chromosome 19.

[0054] Furthermore, the use of the marker in preparing a kit for tissue tracing can be obtained. It can be understood that the marker includes one or two or more of the above markers:

[0055] MST1: genomic region from 49723693 to 49724001 on chromosome 3, HS3ST1: genomic region from 11429523 to 11429690 on chromosome 4, OSMR: genomic region from 38846033 to 38846336 on chromosome 5, PTGER4: genomic region from 40681528 to 40681842 on chromosome 5, IRF4: genomic region from 392623 to 392926 on chromosome 6, REPI N1: the genomic region from 150068516 to 150068824 of chromosome 7, TMEM215: the genomic region from 32783168 to 32783474 of chromosome 9, PAX5: the genomic region from 37027033 to 37027333 of chromosome 9, KRT5: the genomic region from 52913892 to 52914230 of chromosome 12, DNM2: the genomic region from 10823533 to 10823837 of chromosome 19.

[0056] Based on the above markers, a composition for tissue tracing can be obtained, which includes a nucleic acid for detecting the methylation state of the above markers. The methylation state of the above markers can be characterized by the methylation of the target sequence of the above markers.

[0057] Furthermore, a tissue tracing kit can be obtained, which contains reagents for detecting the above-mentioned markers, and can be used to trace the cancer tissue samples of a cancer patient population to help determine the type of cancer they have.

[0058] Step S130, extract the whole genome fragment information of the reference genome, divide the whole genome fragment information into multiple fragment intervals, and calculate the fragment coverage of each fragment interval in the multiple fragment intervals. It can be understood that the whole genome data of the reference genome can be preprocessed, and an example is to filter out the low-quality sequencing reads (reads) with a mapping quality (MAPQ, Mapping Quality) < a specified value, such as a MAPQ value <30 through a bam file, and then use a genome fragment extraction tool, such as an R package GCcontent, to extract the whole genome fragment information; according to the position on the autosome, the extracted fragment information is tiled into adjacent, non-overlapping multiple fragment intervals. In some embodiments, the reference genome can be selected as the human hg19 reference genome, and the length a of the multiple fragment intervals is set to 1Mb.

[0059] Specifically, in each fragment interval, the short fragment is set as a short fragment, and its length can be set between 100-150bp, the middle fragment is set as a middle fragment, and its length is set between 151-260bp, and the long fragment is set as a long fragment, and its length is set between 261-320bp; in addition, the above three fragments can also be integrated to obtain a valid fragment, namely a cov fragment, whose interval length is between 100-320bp. Then, the number of short, middle, long, and cov in each fragment interval is counted. Coverage is defined for each fragment interval: j represents the jth fragment interval among multiple fragment intervals, then the various types of fragments short, middle, long, and cov in the jth fragment interval and their corresponding numbers can be set as fragment features and corresponding feature values. In this way, the elastic network regression algorithm can be used to construct a fragmented binary cancer tissue tracing classification model as a fragmentation model.

[0060] In particular, the purpose of selecting elastic network regression to construct a fragmentation model is that the number of features contained in a fragment may be much higher than the number of fragments, and elastic network regression can effectively select fragment features to avoid overfitting of the fragmentation model. In this way, fragmentation can show better results than the methylation model in dealing with the high collinearity problem between chromosome fragments. This is because there are usually complex interactions and associations between chromosome fragments, and the methylation model constructed using candidate intervals does not take this into account; in addition, elastic network regression is more suitable for constructing sparse models to reduce the complexity and computational complexity of the model. Therefore, the fragmentation model can show higher specificity in classification and has a faster fitting speed. The specificity of the methylation model can be effectively improved by constructing a fragmentation model and integrating it with the methylation model.

[0061] It can be understood that the fragment features are directly obtained through the reference genome, without the need for any type of whole genome sequencing of the cfDNA sample. After obtaining the fragment features, they can be further reduced in dimension. For example, the PCA method is used to reduce the dimension of a large number of fragment features, and a fragment feature matrix is ​​constructed with principal components that can explain 90% of the differences, and a fragmentation model is constructed through an elastic network regression algorithm. The purpose of selecting elastic network regression is that the number of fragment features in the fragment interval is much higher than the number of fragment intervals, and elastic network regression can effectively select fragment features to avoid overfitting of the fragmentation model.

[0062] Step S140, construct a methylation model through multiple candidate intervals, construct a fragmentation model through fragment coverage, and integrate the methylation model and the fragmentation model to obtain a joint traceability model. In a favorable embodiment, cfDNA samples of various cancer patients are input into the methylation model and the fragmentation model to obtain the tissue traceability prediction values ​​of the cancer patients, respectively. It is possible to choose to use a logistic regression algorithm to train a binary multimodal pan-cancer early prediction model through the tissue traceability prediction values ​​of the cancer patients, and predict the tissues of various cancer patients separately. It should be noted that after obtaining the tissue traceability prediction values ​​of the cancer patients through steps S110-S130, that is, after at least one model predicts the tissue labels of the cancer patients, it is also possible to obtain a prediction model of three or more classification items as a joint traceability model through any feasible algorithm training, or to use the structured or unstructured text of the tissue prediction values ​​of the cancer patients to fine-tune the large language model to obtain the multi-classification results of tissue traceability, and this application does not limit this.

[0063] Exemplary Devices

[0064] Figure 4 The figure illustrates a block diagram of a tissue tracing device based on cfDNA methylation sequencing according to an embodiment of the present application.

[0065] like Figure 4As shown, the tissue tracing device 200 based on cfDNA methylation sequencing according to an embodiment of the present application includes:

[0066] The sample sequencing unit 210 collects cfDNA samples from various cancer patients and healthy people, and performs targeted methylation sequencing on the cfDNA samples to obtain multiple methylation intervals; the interval screening unit 220 calculates the significance value of each methylation interval in the multiple methylation intervals, deletes the corresponding methylation interval in response to the significance value being not less than a first threshold, retains the corresponding methylation interval in response to the significance value being less than the first threshold, and screens the multiple methylation intervals to obtain candidate intervals; the fragment extraction unit 230 extracts the whole genome fragment information of the cfDNA sample, divides it into multiple fragment intervals according to the reference genome, and calculates the fragment coverage of each fragment interval in the multiple fragment intervals; the model construction unit 240 constructs a methylation model through multiple candidate intervals, constructs a fragmentation model through fragment coverage, and integrates the methylation model and the fragmentation model to obtain a joint tracing model for cancer tissue tracing of various cancer patients.

[0067] Here, those skilled in the art can understand that the specific functions and operations of each unit and module in the above-mentioned tissue tracing device 200 based on cfDNA methylation sequencing have been referred to above. Figure 1 to Figure 2 The description of the tissue tracing method based on cfDNA methylation sequencing has been introduced in detail, and therefore, its repeated description will be omitted.

[0068] As described above, the tissue tracing device 200 based on cfDNA methylation sequencing according to the embodiment of the present application can be implemented in various terminal devices, such as servers for training methylation models, fragmentation models, and joint tracing models. In one example, the tissue tracing device 200 based on cfDNA methylation sequencing according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the tissue tracing device 200 based on cfDNA methylation sequencing can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the tissue tracing device 200 based on cfDNA methylation sequencing can also be one of the many hardware modules of the terminal device.

[0069] Alternatively, in another example, the tissue tracing device 200 based on cfDNA methylation sequencing and the terminal device may also be separate devices, and the tissue tracing device 200 based on cfDNA methylation sequencing may be connected to the terminal device via a wired and / or wireless network, and transmit interactive information in accordance with an agreed data format.

[0070] Exemplary Electronic Devices

[0071] Below, reference Figure 5 To describe an electronic device according to an embodiment of the present application.

[0072] Figure 5 A block diagram of an electronic device according to an embodiment of the present application is illustrated.

[0073] like Figure 5 As shown, the electronic device 10 includes one or more processors 11 and a memory 12 .

[0074] The processor 13 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0075] The memory 12 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on a computer-readable storage medium, and the processor 11 may run the program instructions to implement the tissue tracing method based on cfDNA methylation sequencing of the various embodiments of the present application described above and / or other desired functions. Various contents such as candidate intervals, joint tracing models, etc. may also be stored in the computer-readable storage medium.

[0076] In one example, the electronic device 10 may further include: an input device 13 and an output device 14, and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0077] The input device 13 may include, for example, a keyboard, a mouse, etc.

[0078] The output device 14 can output various information to the outside, including the trained joint traceability model, etc. The output device 14 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0079] Of course, to simplify, Figure 5 Only some of the components related to the present application in the electronic device 10 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device 10 may also include any other appropriate components.

[0080] Example

[0081] This application provides a general and / or specific description of the materials and test methods used in the experiment. The raw materials or instruments used without indicating the manufacturer are all conventional raw materials or instruments that can be obtained through commercial purchase.

[0082] Example 1 Model feature engineering

[0083] We collected 10-20 ml of peripheral blood from 51 lung cancer, 67 colorectal cancer, 46 gastric cancer, 31 liver cancer, 30 esophageal cancer, 26 thyroid cancer, and 37 ovarian cancer subjects from Boercheng, separated the plasma samples, and extracted and purified cfDNA. Panel capture and methylation library sequencing were then performed. The data of 34 lung cancer, 46 colorectal cancer, 34 gastric cancer, 21 liver cancer, 20 esophageal cancer, 18 thyroid cancer, and 24 ovarian cancer subjects were used as training sets, and the rest were used as test sets.

[0084] 1. Screening of methylation intervals

[0085] 1) Interval definition: MethBin, which is the interval adjacent to the CpG site obtained after merging sequencing (interval length <400bp, interval contains ≥3 CpG sites).

[0086] 2) In the training set, the Kruskal-Wallis test was first used to obtain the p-value of each interval, and the p-value was corrected using the FDR method to obtain the qvalue as the significance value, and the methylation intervals with a qvalue less than 0.05 were retained as candidate intervals.

[0087] 3) For the candidate intervals, calculate the methylation value difference Delta between the seven cancers, for example, the difference between lung cancer and the other six cancers, recorded as Lung_Colorectum_Delta, Lung_Stomach_Delta, Lung_Liver_Delta, Lung_Esophagus_Delta, Lung_Thyroid_Delta and Lung_Ovary_Delta. For each cancer, calculate the difference between it and the other six cancer types. Taking lung cancer as an example, for a candidate interval, if 4 or more of the above six Delta values ​​are greater than 0.05, we will retain the interval. Perform the same operation for other cancers, and finally we can obtain the interval after further filtering for each cancer type.

[0088] 4) For the candidate intervals, the improved random forest feature screening algorithm of the R package Boruta is used to calculate the feature importance value of the candidate interval.

[0089] 2. Fragmentation feature extraction

[0090] Through the bam file, reads with MAPQ < 30 were filtered out, and the whole genome fragment information of the reference genome was extracted using the R package GCcontent. The extracted fragment information was tiled into adjacent, non-overlapping 1Mb intervals according to the hg19 reference genome autosomes. Short fragments are defined as short, between 100-150bp, medium fragments are defined as middle, between 151-260bp, long fragments are defined as long, between 261-320bp, and cov is a combination of short, middle and long fragments. The number of short, middle, long and cov in each interval is counted, that is, the fragment coverage. Finally, the PCA algorithm is used to calculate the importance value of all fragment coverage, sort and select the top principal components that can explain 90%.

[0091] Example 2 Construction and evaluation of methylation model

[0092] For the candidate intervals finally screened in Example 1, the top 10 methylation intervals with the highest feature importance values ​​were selected, and the tissue tracing model for the above 7 cancers was constructed using the elastic regression network in the training set and tested in the test set. The performance of the model in the training set and the test set is as follows: Figure 3A The model has a top1 accuracy of 89% and a top2 accuracy of 75.8% on the test set.

[0093] Example 3 Construction and evaluation of fragmentation model

[0094] For the 90% principal components obtained by random forest in Example 1, an elastic regression network was trained to construct the tissue tracing model of the above 7 cancers, and the model was tested on the test set. The performance of the model on the training set and the test set is as follows: Figure 3B The model has a top1 accuracy of 87.5% and a top2 accuracy of 75.8% on the test set.

[0095] Example 4 Construction and evaluation of a joint traceability model based on a methylation model and a fragmentation model

[0096] For the models in Example 2 and Example 3, the score values ​​of each cfDNA sample for the above 7 types of cancer, i.e., the cancer tissue prediction value, can be obtained respectively. The score values ​​obtained by the two models are used as features, and a new joint traceability model is constructed using the logistic regression method. The prediction is performed on the test set. The performance of the model in the training set and the test set is shown in FIG. Figure 3C The model has a top1 accuracy of 91.2% and a top2 accuracy of 80.2% on the test set.

[0097] In combination with Examples 1 to 4, the tissue tracing method based on cfDNA methylation sequencing adopted in the present application is used to predict the tissue origin of cfDNA samples of the target population. The method has high accuracy and good sample detection sensitivity, and the prediction accuracy of multiple cancer tissues is above 80%.

[0098] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0099] In this application, words such as "including", "comprising", "having", etc. are open words, meaning "including but not limited to", and can be used interchangeably therewith. The words "or" and "and" used herein mean the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein means the phrase "such as but not limited to", and can be used interchangeably therewith.

[0100] It should also be noted that in the method, system and device of the present application, each step or each module can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0101] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A tissue tracing method based on cfDNA methylation sequencing, comprising: Collect cfDNA samples from various cancer patients and healthy people and obtain multiple methylation intervals through methylation sequencing; Calculating a significance value of each methylation interval in the plurality of methylation intervals, deleting the corresponding methylation interval in response to the significance value being not lower than a first threshold, retaining the corresponding methylation interval in response to the significance value being lower than the first threshold, and screening the plurality of methylation intervals to obtain a plurality of candidate intervals; Extracting whole genome fragment information based on the reference genome and equally dividing it into multiple fragment intervals, and calculating the fragment coverage of each fragment interval in the multiple fragment intervals; A methylation model is constructed using the multiple candidate intervals, a fragmentation model is constructed using the fragment coverage, and the methylation model and the fragmentation model are integrated to obtain a joint tracing model for cancer tissue tracing in various cancer populations.

2. The tissue tracing method based on cfDNA methylation sequencing according to claim 1, wherein: The screening of the multiple methylation intervals to obtain multiple candidate intervals comprises: respectively calculating the methylation value of each methylation interval in the plurality of methylation intervals for each cancer patient population in the plurality of cancer patient populations; The difference between the methylation value of each methylation interval in each cancer population and the methylation value in other cancer populations among the multiple cancer populations is calculated respectively; in response to the number of the methylation value differences greater than the second threshold being not less than the third threshold, the corresponding methylation interval is retained; in response to the number of the methylation value differences greater than the second threshold being less than the third threshold, the corresponding methylation interval is deleted.

3. The tissue tracing method based on cfDNA methylation sequencing according to claim 2, wherein: The step of respectively calculating the methylation value of each methylation interval in the plurality of methylation intervals for each cancer patient population in the plurality of cancer patient populations comprises: The average methylation value of the CpG site in each methylation interval of the multiple methylation intervals in each cancer patient population among the multiple cancer patient populations is calculated respectively.

4. The tissue tracing method based on cfDNA methylation sequencing according to claim 1, wherein: The screening of the multiple methylation intervals to obtain candidate intervals comprises: respectively calculating the importance value of each methylation interval in the plurality of methylation intervals; sorting the multiple methylation intervals according to the importance values; The first n methylation intervals among the multiple methylation intervals are retained, and the remaining methylation intervals are deleted.

5. The tissue tracing method based on cfDNA methylation sequencing according to claim 1, wherein: The step of extracting whole genome segment information based on the reference genome and dividing the information into multiple segment intervals comprises: The whole genome fragment information is flattened according to the reference genome autosomes and divided into the plurality of adjacent fragment intervals with an interval length of a.

6. The tissue tracing method based on cfDNA methylation sequencing according to claim 5, wherein: The calculating the segment coverage of each segment interval in the plurality of segment intervals comprises: Extracting short segments, medium segments, long segments and valid segments from the multiple segment intervals, and recording the number of the short segments, the medium segments, the long segments and the valid segments; wherein, The interval length of the short fragment is 100-150 bp; The length of the middle fragment is 151-260 bp; The long fragment has a length of 261-320 bp; and The effective fragment has an interval length of 100-320 bp.

7. The tissue tracing method based on cfDNA methylation sequencing according to claim 1, wherein: The integrating the methylation model and the fragmentation model to obtain a joint traceability model comprises: The methylation model and the fragmentation model are used to obtain the tissue tracing prediction value of the cfDNA sample of each cancer patient population in multiple cancer patient populations; Based on the tissue traceability prediction value, a joint traceability model is trained through logistic regression.

8. A tissue tracing device based on cfDNA methylation sequencing, comprising: The sample sequencing unit collects cfDNA samples from a variety of cancer patients and healthy people and obtains multiple methylation intervals through methylation sequencing; an interval screening unit, calculating a significance value of each methylation interval in the plurality of methylation intervals, deleting the corresponding methylation interval in response to the significance value being not lower than a first threshold, retaining the corresponding methylation interval in response to the significance value being lower than the first threshold, and screening the plurality of methylation intervals to obtain a plurality of candidate intervals; A fragment extraction unit, which extracts whole genome fragment information based on a reference genome and equally divides the fragment information into multiple fragment intervals, and calculates the fragment coverage of each of the multiple fragment intervals; A model building unit builds a methylation model through the multiple candidate intervals, builds a fragmentation model through the fragment coverage, and integrates the methylation model and the fragmentation model to obtain a joint tracing model for cancer tissue tracing of various cancer patients.

9. The tissue tracing device based on cfDNA methylation sequencing according to claim 8, which executes the tissue tracing method based on cfDNA methylation sequencing according to any one of claims 1-7.

10. An electronic device comprising: processor; as well as A memory, in which computer program instructions are stored, and when the computer program instructions are executed by the processor, the processor executes the tissue tracing method based on cfDNA methylation sequencing as described in any one of claims 1-7.

Images