A method for identifying regions of abnormally high methylation in cancer methylation data

Abstract

The application discloses a method for identifying abnormal hypermethylation regions in cancer methylation data, comprising: obtaining methylation sequencing data of normal tissues and cancer tissues; identifying reference UMRs based on normal tissue methylation data; screening abnormal hypermethylation regions from the methylation sequencing data of the normal tissues and the cancer tissues based on the reference UMRs, a hidden Markov model and hypothesis testing; calculating the proportion and / or length of hypermethylated CpG sites in the abnormal hypermethylation regions relative to CpG sites in the reference UMRs; classifying the abnormal hypermethylation regions into phUMRs or fhUMRs based on the proportion and / or length; the phUMRs represent that CpG hypermethylation only occurs in part of the UMRs, and the fhUMRs represent that hypermethylation occurs in the whole UMRs.

Description

Technical Field

[0001] This invention relates to the field of biological data analysis, and more specifically, to a method and system for identifying abnormally hypermethylated regions in cancer methylation data. Background Technology

[0002] Hotspot mutations in the IDH1 and IDH2 genes are frequently found in malignant gliomas, acute myeloid leukemia, and several other cancers. Studies have shown that IDH hotspot mutations are involved in gliomas as an early driver of diffuse low-grade gliomas and secondary glioblastomas. The mutated IDH gene does not produce α-ketoglutarate but instead produces D-2-hydroxyglutarate (D2HG), which competitively inhibits iron-dependent hydroxylases, including TET family enzymes that mediate active DNA demethylation. Therefore, aberrant methylation patterns in undermethylated regions, such as the glioma-CpG island methylsomal phenotype (G-CIMP), are observed in IDH-mutant gliomas.

[0003] Although previous studies have shown that IDH mutations are sufficient to stabilize aberrant methylation in gliomas, the role of these altered methylation states and their relationship with glioma chromatin characteristics remain unclear. Limited by probe design biases in methylation microarrays, previous studies on aberrant methylation in IDH-mutant gliomas have primarily focused on CPG islands or promoters. Furthermore, these studies are based on the average methylation level of all CpG sites in each genomic region; therefore, when methylation alterations occur only in CpG islands or partial promoters, these analytical methods cannot provide accurate quantitative analysis. To fully explain the chromatin reprogramming and functional effects of these altered methylation states, accurate quantification of methylation changes at individual CpG site resolution is essential. Summary of the Invention

[0004] This invention aims to address at least one of the technical problems existing in the prior art. To this end, this invention provides a method and system for identifying anomalously hypermethylated regions in cancer methylation data. The method of this invention, by developing a method for identifying anomalously hypermethylated regions and disclosing the process of classifying the identified anomalously hypermethylated regions, provides a computational strategy for accurately decoding methylation erosion patterns in IDH-mutant gliomas, revealing their potential mechanistic insights into chromatin reprogramming during tumorigenesis.

[0005] The first aspect of this application discloses a method for identifying abnormally hypermethylated regions in cancer methylation data, comprising: acquiring methylation sequencing data of normal tissue and cancer tissue;

[0006] Abnormally hypermethylated regions were screened from the methylation sequencing data of the normal and cancer tissues based on reference UMRs.

[0007] Calculate the ratio of hypermethylated CpG sites in the anomalously hypermethylated region to CpG sites in the reference UMRs;

[0008] Based on the aforementioned ratio, the abnormally hypermethylated regions are divided into phUMRs or fhUMRs; phUMRs indicate that CpG hypermethylation occurs only in a portion of the UMRs, while fhUMRs indicate that hypermethylation occurs throughout the entire UMRs.

[0009] The calculation of the proportion of hypermethylated CpG sites in the abnormally hypermethylated region includes:

[0010] The hypermethylation state within the abnormally hypermethylated region is determined using machine learning algorithms and / or statistical methods; the CpG sites within the hypermethylated state are the hypermethylated CpG sites.

[0011] The CpG sites in the Hyper methylation state are merged into the methylation region to obtain the methylation region after merging CpG sites, which is the abnormally hypermethylated region.

[0012] Calculate the proportion of individual hypermethylated CpG sites in the methylation region after merging CpG sites;

[0013] Optionally, when determining the methylation state using machine learning algorithms, the method further includes modeling the emission probability matrix of the methylation state using mathematical distribution methods, and calculating the mean and variance based on the methylation differences between normal and cancer tissue sequencing data (calculating the methylation differences of all CpG sites between normal and cancer sites, and then calculating the mean and variance); calculating the methylation levels of adjacent CpG sites to obtain the transition probability (the transition probability is an internal parameter of the Hidden Markov Model (HMM). This is equivalent to providing training parameters for the model. After the model is trained, it is used to find which sites are Hyper).

[0014] Optionally, when determining the methylation status using statistical methods, a methylation matrix for each CpG site in each of the reference UMRs is established; the p-value for each CpG site in the reference UMRs is calculated using a two-tailed test, and the state of CpG sites with an FDR-corrected p-value less than 0.05 and an absolute methylation difference greater than a first threshold (0.2) is Hyper.

[0015] Optionally, the machine learning algorithm includes one or more of the following: Hidden Markov Model, Hotspot Augmentation Algorithm, Sliding Window Algorithm, preferably Hidden Markov Model (HMM);

[0016] Optionally, the mathematical distribution method includes one or more of the following: Gaussian distribution, beta distribution, binomial distribution, preferably Gaussian distribution.

[0017] The CpG sites within the Hypermethylation state include: CpG sites within the Hypermethylation state determined using a machine learning algorithm;

[0018] Alternatively, the CpG sites within the Hyper methylation state can be determined using statistical methods;

[0019] Alternatively, the CpG sites within the Hypermethylation state determined by machine learning algorithms and the CpG sites within the Hypermethylation state determined by statistical methods can be combined to obtain the CpG sites after taking the union or intersection.

[0020] Optionally, the methylation sequencing data is WGBS data.

[0021] The step of classifying the abnormally hypermethylated regions into phUMRs or fhUMRs based on the ratio includes: the ratio being less than or equal to or less than a second threshold (0.8) for phUMRs, and the ratio being greater than or equal to or greater than the second threshold for fhUMRs.

[0022] Optionally, the method further includes: after screening out abnormally hypermethylated regions from the methylation sequencing data of the normal tissue and cancer tissue based on reference UMRs, calculating the length of the hypermethylated CpG sites in the abnormally hypermethylated regions, and classifying the abnormally hypermethylated regions into phUMRs or fhUMRs based on the length; the length being greater than or equal to or greater than a third threshold is phUMRs, and the length being less than or equal to or less than a second threshold is fhUMRs;

[0023] Optionally, the method of classifying the abnormally hypermethylated regions into phUMRs or fhUMRs further includes: classifying the abnormally hypermethylated regions into phUMRs or fhUMRs based on the ratio and length; wherein the ratio is less than or equal to or less than a second threshold, and the length (number of CGs) is greater than or equal to or greater than a third threshold, and the ratio is greater than or equal to the second threshold, in which case they are phUMRs.

[0024] The results of phUMR typing were used to determine the activation of oncogenes in the sample.

[0025] Optionally, based on the typing results of the phUMRs, one or more of the following results can also be obtained in the sample: the genes in the phUMRs on the promoter are upregulated, the upregulated genes are rich in oncogenes, the upregulated genes on the promoter have increased activity histone modification signals (H3K4me3, H3K27ac, H3K4me1) and decreased repressive histone (H3K27me3) signals.

[0026] The method further includes defining highly methylated regions within the phUMRs as partially hypermethylated regions compared to the reference UMRs, and defining low-methylated regions within the phUMRs as flanking UMRs. That is, each phUMR consists of two parts: partially hypermethylated and flanking UMR.

[0027] The method further includes: based on the results of partiallyHyper, obtaining results showing significant enrichment of H3K4me3 and H3K27ac (tumor suppressor genes), and a reduction in related phUMR activity modification signals for promoter upregulation and downregulation; based on the results of flankingUMR, obtaining results showing an increase in phUMR activity modification signals for promoter upregulation.

[0028] A second aspect of this application discloses a device for identifying abnormally hypermethylated regions in cancer methylation data, the device comprising: a memory and a processor;

[0029] The memory is used to store program instructions; the processor is used to invoke the program instructions, and when the program instructions are executed, they are used to perform the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data as described in the first aspect of this application.

[0030] The third aspect of this application discloses a computer program product, including a computer program that, when executed by a processor, implements the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data as described in the first aspect of this application.

[0031] The fourth aspect of this application discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data as described in the first aspect of this application.

[0032] This application has the following beneficial effects:

[0033] 1. This application innovatively discloses a computational framework based on a Hidden Markov Model to identify single-base resolution hypermethylated UMR in IDH-mutant gliomas. This hypermethylated UMR (hypermethylation state) is a finer-grained hypermethylated region compared to conventional methods. Compared to traditional methods that calculate the average methylation level of all CpG sites in CpG islands, this application uses a reference UMR as a benchmark to more comprehensively reflect undermethylated regions in normal brain tissue, precisely quantifying the UMR obtained from analyzing individual CpG sites and revealing methylation changes. Using the average methylation level within the hypermethylated region can quantify methylation differences better than calculating the average methylation level of all CpG sites within the UMR.

[0034] 2. This application innovatively further studies the methylation status in hypermethylated UMRs, classifying them into partially hypermethylated UMRs (phUMRs) and fully hypermethylated UMRs (fhUMRs) based on their bimodal methylation. Furthermore, it reveals that phUMRs exhibit different genomic and histone characteristics compared to fhUMRs, and that phUMRs are associated with the activation of oncogenes in the samples. This method delves deeper into the life principles hidden behind biological data, significantly improving the accuracy and depth of data analysis.

[0035] 3. This application innovatively conducts further research on phUMRs, dividing them into partially hypermethylated regions and flanyingUMRs with internal hypomethylation, enabling more precise and targeted research into the relationship between methylation status and disease. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic flowchart of a method for identifying abnormally hypermethylated regions in cancer methylation data provided by an embodiment of the present invention;

[0038] Figure 2 This is a schematic diagram of a device for identifying abnormally hypermethylated regions in cancer methylation data provided in an embodiment of the present invention;

[0039] Figure 3This is a schematic flowchart of a system for identifying abnormally hypermethylated regions in cancer methylation data provided in an embodiment of the present invention;

[0040] Figure 4 This is a diagram showing the identification results of phUMRs provided in an embodiment of the present invention;

[0041] Figure 5 This is a schematic diagram illustrating the features of phUMRs and fhUMRs provided in the embodiments of the present invention;

[0042] Figure 6 This is a transcriptional pattern diagram of phUMRs and fhUMRs-related genes provided in the embodiments of the present invention. Detailed Implementation

[0043] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0044] In some of the processes described in the specification, claims, and accompanying drawings of this invention, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to different types.

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Figure 1 This is a schematic flowchart illustrating a method for identifying abnormally hypermethylated regions in cancer methylation data according to an embodiment of the present invention. Specifically, the method includes the following steps:

[0047] 101: Obtain methylation sequencing data from normal and cancerous tissues;

[0048] In one embodiment, the methylation sequencing data of cancer tissue is WGBS data from IDH-mutant glioma tissue. Whole-genome bisulfite sequencing (WGBS) treats genomic DNA with sodium bisulfite to convert unmethylated cytosine C to uracil U, which becomes T in subsequent PCR and sequencing processes, while methylated C remains unaffected. By resequencing the treated DNA and comparing it with a reference genome, high-precision methylation level analysis at single-base resolution is achieved at the genomic level.

[0049] In one embodiment, the protein encoded by the IDH1 gene is called isocitrate dehydrogenase 1. There are three types of human IDH: IDH1, IDH2, and IDH3. IDH1 is located in the cytoplasm and peroxisomes, while IDH2 and IDH3 are located in mitochondria. This type of protease oxidizes isocitrate to oxaloylsuccinate, which is then converted to α-ketoglutarate. Initial studies found a close association between IDH1 mutations and gliomas; later, mutations were found to be associated with prostate cancer, paragangliomas, and IDH1 / 2 mutations with acute myeloid leukemia. The tumorigenic mechanism involves the mutation of IDH converting α-ketoglutarate to 2-hydroxyglutarate, which inhibits the former's targets, leading to abnormal expression of these targets and inducing cancer.

[0050] 102: Based on reference UMRs, abnormally hypermethylated regions were screened from the methylation sequencing data of the normal and cancerous tissues;

[0051] In one embodiment, the method for obtaining the reference UMRs as under-methylated regions (refUMRs) includes:

[0052] Acquire methylation sequencing data; (preprocessing, using BSMAP to trim adapters, low-quality and repetitive sequences with a default threshold); process the methylation sequencing data using existing software to obtain the methylation ratio of each CpG covered by at least N (N is an integer greater than 1, preferably 4) reads; determine UMRs from at least M consecutive low-methylated CpGs; merge overlapping UMRs from multiple samples and remove redundant UMRs; calculate the UMR frequency (UOF) of intersecting fragments from multiple samples; identify UMRs with high UOF as the reference UMRs. Specifically, this embodiment provides a comprehensive statistical framework to identify human reference UMRs from 65 high-quality WGBS profiles (whole genome CPG coverage >90%), mainly including the following four steps:

[0053] Step 1: For each WGBS matrix, use BSMAP to trim transitions, low-quality, and repetitive sequences with default thresholds, and align sodium bisulfite-treated reads with the human genome (hg19). Use coverage of 4 reads to ensure the accuracy of CpG methylation detection. Calculate the methylation ratio of each CPG covered by at least 4 reads using the percentage value (bsratio) module in BSMAP. BSMAP is a well-known methylation sequencing alignment software that uses a "wild-card" strategy for alignment. Simply put, it creates a "SeedTable" using the reference genome. Each C in each seed can be either methylated or unmethylated, so the "SeedTable" contains all possible C->T conversions. Reads are then aligned with the "SeedTable," and the optimal alignment is selected. Bismark is another well-known methylation sequencing alignment software. Before alignment, Bismark performs C->T and G->A conversions on all reads and also performs these two conversions on the reference genome. For non-strand-specific library construction reads whose orientation is uncertain, the two converted reads must be compared separately with two different scenarios of the reference genome. This is equivalent to performing four different comparisons for each read, from which the best alignment is selected as the result. If strand-specific library construction is used, since the strand affiliation of the reads is known, only two scenarios are compared, which is much faster. Because Bismark performs C->T or G->A conversions on the reads and the reference genome, in one case only three bases remain after the conversion, the working principle of Bismark is commonly referred to as "three-base alignment."

[0054] Step 2: UMRs were determined from at least four consecutive hypomethylated CpGs (hypo-average methylation rate of 10%). To reduce the impact of sparse CPG density on UMR detection in the HMM-based model, UMRs with an Obs / Exp value < 0.1 were removed.

[0055] Step 3: Redundant UMRSs were reduced from multiple tissue and tumor WGBS profiles by merging overlapping UMRs from multiple samples (35,217,985 in total). To characterize the enrichment distribution of genome-wide UMRs in tissue and tumor samples, the UMR frequency (UOF) of crossover segments in N samples was calculated as follows:

[0056]

[0057]

[0058] The UMR proportion score represents the proportion of UMR in population-sized samples of normal tissue and tumor. A higher UOF indicates more conserved UMR in population-sized samples. Conversely, a decrease in UOF represents a shortening or loss of UMR, suggesting hypermethylation in these regions at the population scale.

[0059] Step 4: Reference UMRs were detected from the UOF spectra of the sample population using the Poisson test (p-value < 1.0e-8, p-value adjusted using the Benjamini and Hochberg (BH) method). These reference UMRs were identified from normal tissues and tumors, respectively.

[0060] In one embodiment, WGBS data from 72 normal brain tissues and 15 IDH-mutant gliomas were analyzed. 21,716 reference UMRs were identified from the 72 normal brain tissues, of which 2,831 reference UMRs were abnormally hypermethylated in the IDH-mutant gliomas. The process involved first identifying the reference UMRs, then identifying the UMRs of the tumor samples, and comparing the two to obtain cancer-related hypermethylated UMRs.

[0061] In one embodiment, after acquiring WGBS data from normal brain tissue and IDH-mutant gliomas, the two WGBS datasets are merged. The merging process includes: at each CpG site, merging reads on the positive and negative strands to improve read coverage; only CpG sites with more than five reads are considered for analysis, and CpG sites on the scaffold are removed; principal component analysis (PCA) of CpG island methylation levels is used to assess the concordance between WGBS samples from different sources.

[0062] 103: Calculate the proportion of hypermethylated CpG sites in the anomalously hypermethylated region relative to the CpG sites in the reference UMRs;

[0063] In one embodiment, to explore anomalous methylation changes in abnormally hypermethylated regions, this embodiment detects the proportion of hypermethylated CpG sites in undermethylated regions. Interestingly, not all hypermethylated UMRs are fully hypermethylated; a bimodal pattern is observed, such as... Figure 4 As shown in B and C; the calculation of the proportion of hypermethylated CpG sites in the abnormally hypermethylated regions includes:

[0064] The hypermethylation state within the anomalously hypermethylated region is determined using machine learning algorithms and / or statistical methods; the CpG site within the hypermethylation state is the hypermethylated CpG site; optionally, the methylation state further includes: hypermethylated or no diff.

[0065] The CpG sites in the Hyper methylation state are merged into the methylation region to obtain the methylation region after merging CpG sites, which is the abnormally hypermethylated region.

[0066] Based on the bimodal distribution curve of the proportion of hypermethylated CpG sites in the reference UMRs, the proportion of hypermethylated CpG sites in the methylated region after merging CpG sites is calculated.

[0067] Optionally, when determining the methylation state using machine learning algorithms, the method further includes modeling the emission probability matrix of the methylation state using mathematical distribution methods, and calculating the mean and variance based on the methylation differences between normal and cancer tissue methylation sequencing data (calculating the methylation differences of all CpG sites between normal and cancer sites, and then calculating the mean and variance); calculating the methylation levels of adjacent CpG sites to obtain the transition probability (the transition probability is an internal parameter of the Hidden Markov Model (HMM). This is equivalent to providing training parameters for the model. After the model is trained, it is used to identify which sites are hyper-methylated); for each refUMR, the initial differential methylation state of the first CpG site is set by calculating the average methylation level of that region. RHmm (version 2.0.2) is used to assign different states to each CpG site.

[0068] Optionally, when determining the methylation status using statistical methods, a methylation matrix for each CpG site in each of the reference UMRs is established; the p-value for each CpG site in the reference UMRs is calculated using a two-tailed test, and the status of CpG sites with an FDR-corrected p-value less than 0.05 and an absolute methylation difference greater than a first threshold (0.2) is Hyper.

[0069] Optionally, the machine learning algorithm includes one or more of the following: Hidden Markov Model, Hotspot Augmentation Algorithm, Sliding Window Algorithm, preferably Hidden Markov Model (HMM);

[0070] Optionally, the mathematical distribution method includes one or more of the following: Gaussian distribution, beta distribution, binomial distribution, preferably Gaussian distribution.

[0071] In one embodiment, the CpG sites within the Hypermethylation state include: CpG sites within the Hypermethylation state determined using a machine learning algorithm;

[0072] Alternatively, the CpG sites within the Hyper methylation state can be determined using statistical methods;

[0073] Alternatively, the CpG sites within the Hypermethylation state determined by machine learning algorithms and the CpG sites within the Hypermethylation state determined by statistical methods can be taken as the union or intersection to obtain the CpG sites after taking the union or intersection.

[0074] 104: Based on the stated ratio, the abnormally hypermethylated regions are divided into phUMRs or fhUMRs; phUMRs indicate that CpG hypermethylation occurs only in a portion of the UMR, and fhUMRs indicate that hypermethylation occurs throughout the entire UMR.

[0075] In one embodiment, classifying the abnormally hypermethylated regions into phUMRs or fhUMRs based on the ratio includes: phUMRs having a ratio less than or equal to a second threshold (0.8), and fhUMRs having a ratio greater than or equal to the second threshold. phUMRs indicate that CpG hypermethylation occurs only in a portion of the UMRs, while fhUMRs indicate that hypermethylation occurs throughout the entire UMR. Specifically, as in Example 1, for a reference UMR containing 20 CpG sites, 12 of which are hypomethylated in normal tissue and hypermethylated in tumor tissue, and the remaining 8 are hypomethylated in both normal and tumor tissue, then the proportion of hypermethylated CpG sites in this UMR is 0.6 (12 / 20).

[0076] Optionally, after screening out abnormally hypermethylated regions from the methylation sequencing data of the normal and cancer tissues based on reference UMRs, the method further includes: calculating the length of hypermethylated CpG sites in the abnormally hypermethylated regions, and classifying the abnormally hypermethylated regions into phUMRs or fhUMRs based on the length; the length is greater than or equal to or greater than a third threshold and is classified as phUMRs, and the length is less than or equal to or less than a second threshold and is classified as fhUMRs; specifically, as in Example 2, the phUMR definition includes two thresholds: ① a proportion exceeding 0.8 ② more than 5 hypermethylated CpG sites.

[0077] Optionally, the method of classifying the abnormally hypermethylated regions into phUMRs or fhUMRs further includes: classifying the abnormally hypermethylated regions into phUMRs or fhUMRs based on the ratio and length; wherein the ratio is less than or equal to or less than a second threshold, and the length (number of CGs) is greater than or equal to or greater than a third threshold, and the ratio is greater than or equal to the second threshold, in which case they are phUMRs.

[0078] Based on the phUMR typing results, the activation of oncogenes in the target body can be obtained. Optionally, based on the phUMR typing results, one or more of the following results can also be obtained in the sample: upregulation of genes within phUMRs at the promoter, upregulated genes being rich in oncogenes, increased activity of upregulated gene histone modification signals (H3K4me3, H3K27ac, H3K4me1) and decreased repressive histone (H3K27me3) signals at the promoter. The sample is a clinical sample to be tested.

[0079] In one embodiment, phUMRs appear to be more susceptible to conventional average methylation level methods compared to fhUMRs, such as Figure 4 As shown in D. The phUMRs represent CpG hypermethylation occurring only in parts of the UMR, such as the methylation state of the synaptic protein SYT6 promoter, which is the opposite of the disease mechanism of fhUMRs and represents a novel disease mechanism. The fhUMRs represent hypermethylation occurring throughout the entire UMR, such as the methylation state of the cerebrospinal fluid leak-related gene TPPP3 promoter. Figure 4 As shown in C. Figure 4 This embodiment presents the identification results of phUMRs. A) A computational framework for identifying hypermethylated regions from normal and tumor samples. B) A bimodal distribution of the proportion of hypermethylated CpG sites in hypomethylated regions of IDH-mutant gliomas. C) A genome browser visualization of representative phUMRs (SYT6) and fhUMRs (TPPP3) in IDH-mutant gliomas, showing regions with persistent hypomethylation across multiple brain tissues. The bottom image represents the methylation level trajectory and corresponding expression signals over a larger genome-wide scale. D) A comparison of methylation changes calculated using hypermethylated regions (y-axis) against the entire reference hypomethylated region (x-axis). Methylation changes for each absolute difference in the mean methylation level of all CpG sites between IDH-mutant gliomas and normal brain samples. The commonly used methylation difference threshold of 0.2 is represented by the vertical line in the figure. E) Comparison of methylation levels of SYT6 and TPPP3 using different benchmarks (normal brain tissue n=75, IDH-mutant glioma tissue n=15). Hyper refers to the hypermethylated region in phUMRs or fhUMRs.

[0080] Furthermore, the phUMRs are longer and have higher overlap with promoters and CpG islands, exhibiting strong active histone modification signals (H3K4me3, H3K27ac), and genes within the phUMRs on the promoter are easily upregulated; such as Figure 6 As shown in Figure A, the phUMRs on the promoter exhibit different transcriptional regulatory patterns compared to the classic pattern of fhUMRs on the promoter. Upregulated genes on the promoter show increased activity of histone modification signals (H3K4me3, H3K27ac, H3K4me1) and decreased repressive histone signals (H3K27me3), which may be related to the inhibition of transcriptional repressors. Upregulated genes on the promoter are significantly enriched in neural system development and cell differentiation, indicating the key regulatory role of some hypermethylated regions in cell fate determination. Downregulated genes are enriched in synaptic signal transduction and chemical synaptic transmission, such as... Figure 6 As shown in B, genes on the gene body are related to the negative regulation of biosynthetic processes; upregulated genes tend to enrich methylation-sensitive repressive motifs; upregulated genes associated with phUMRs are rich in oncogenes (partial erosion of oncogene promoters is associated with increased transcription accompanied by local changes in H3K4me3 and downstream changes in H3K36me3); the fhUMRs are intergenic regions and CG-poor regions, with weak active histone modification signals. Figure 6 This embodiment presents a transcriptional pattern diagram of phUMRs and fhUMRs-related genes. Specifically, A) a volcano plot shows the statistical significance (y-axis) and fold change (x-axis) of phUMR and fhUMR-related genes between IDH-mutant gliomas and normal brain tissue (left side, downregulated phUMR / fhUMR-related genes; right side, upregulated phUMRs / fhUMRs-related genes); B) functional annotations of upregulated and downregulated genes in phUMR / fhUMR at the promoter. The p-value was adjusted using the BH method. C) a bar chart shows the statistical significance (y-axis) of KEGG for phUMR and fhUMR genes. The line represents the adjusted p-value threshold of 0.05.

[0081] In one embodiment, the method further includes: defining highly methylated regions within the phUMRs as partially hyper compared to the reference UMRs, and defining low-methylated regions within the phUMRs as flanking UMRs; that is, each phUMR consists of two parts: partially hyper and flanking UMR. Specifically, as Figure 5As shown in E, based on the boundary regions of the abnormally hypermethylated CpG, phUMRs are divided into two categories: partially hyper and flanking UMRs. The partial evolutionary conservation is lower than that of the flanking regions, implying that they occurred later in evolution and may have different regulatory functions in transcriptional switching (e.g., Figure 5 (as shown in F); H3K4me3 and H3K27ac (tumor suppressor genes) were significantly enriched in partially Hyper, with no significant differences on the flanking sides (as shown in F). Figure 5 (As shown in G); upregulation of related phUMR on the promoter increases the activity modification signal on flanking UMR; upregulation and downregulation of related phUMR on the promoter decrease the activity modification signal on partiallyHyper. Figure 5 This is a schematic diagram illustrating the characteristics of phUMRs and fhUMRs provided in this embodiment of the invention; specifically, A) the length distribution of phUMRs and fhUMRs; B) the genomic distribution of phUMRs and fhUMRs; C) the percentage of overlap between phUMRs and fhUMRs and CpG islands, CGI shores, CGI shelves, and open seas; D) the mean signal of histone modifications in phUMRs and fhUMRs within ±3kb in normal brain tissue; E) a graph showing the definition of partial Hyper and flanking UMRs in phUMRs. The boundaries are determined by hypermethylated CpG in normal brain and IDH-mutant glioma tissues; F) evolutionary conservation scores of partial Hyper and flanking UMRs. P-values ​​were tested using a two-tailed t.test; G) the mean signal of histone modifications in phUMRs within ±3kb in normal brain and IDH-mutant glioma tissues. For partial Hyper and flanking UMRs, the mean ChIPseq signal was scaled to 2kb.

[0082] In one embodiment, the method further includes: obtaining results based on the partially Hyper results showing significant enrichment of H3K4me3 and H3K27ac (tumor suppressor genes), and a reduction in related phUMR activity modification signals with promoter upregulation and downregulation; obtaining results based on the flankingUMR results showing an increase in phUMR activity modification signals with promoter upregulation; providing a more precise study of the composition mechanism of phUMRs and laying the groundwork for disease mechanism research.

[0083] Figure 2This invention provides a device for identifying abnormally hypermethylated regions in cancer methylation data. The device includes a memory and a processor. The memory stores program instructions. The processor invokes the program instructions, which, when executed, perform the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data.

[0084] This embodiment also discloses a computer program product, including a computer program that, when executed by a processor, implements the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data as described in the first aspect of this application. Specifically, it includes:

[0085] Figure 3 This invention provides a system for identifying abnormally hypermethylated regions in cancer methylation data, comprising:

[0086] Acquisition unit 301 is used to acquire methylation sequencing data of normal tissue and cancer tissue;

[0087] The first processing unit 302 screens out abnormally hypermethylated regions from the methylation sequencing data of the normal and cancerous tissues based on reference UMRs;

[0088] The second processing unit 303 is used to calculate the ratio of hypermethylated CpG sites in the abnormally hypermethylated region to CpG sites in the reference UMRs.

[0089] The third processing unit 304 divides the abnormally hypermethylated regions into phUMRs or fhUMRs based on the ratio; phUMRs indicate that CpG hypermethylation occurs only in a portion of the UMRs, while fhUMRs indicate that hypermethylation occurs throughout the entire UMRs.

[0090] This embodiment also provides a system for identifying abnormally hypermethylated regions in cancer methylation data, including: modifying the second processing unit to: calculate the ratio of hypermethylated CpG sites in the abnormally hypermethylated region to CpG sites in the reference UMRs, or calculate the length of hypermethylated CpG sites in the abnormally hypermethylated region;

[0091] Correspondingly, the third processing unit is modified to: classify the abnormally hypermethylated regions into phUMRs or fhUMRs based on the ratio and / or length.

[0092] This embodiment also provides a system for identifying abnormally hypermethylated regions in cancer methylation data. After the third processing unit, it further includes a fourth processing unit, which obtains the result of activation of oncogenes in the sample based on the typing results of the phUMRs.

[0093] It also includes: a classification unit that defines highly methylated regions within the phUMRs as partially Hyper compared to the reference UMRs, and hypomethylated regions within the phUMRs as flanking UMRs;

[0094] The fifth processing unit, based on the results of partiallyHyper, obtained significant enrichment of H3K4me3 and H3K27ac (tumor suppressor genes), and reduced activity modification signals of related phUMRs with promoter upregulation and downregulation; based on the results of flankingUMR, it obtained increased activity modification signals of phUMRs with promoter upregulation.

[0095] This embodiment also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method steps for identifying abnormally hypermethylated regions in cancer methylation data.

[0096] The verification results of this verification embodiment show that assigning inherent weights to indications can moderately improve the performance of this method compared to the default settings.

[0097] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0098] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0099] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0100] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0101] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0102] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0103] The computer device provided by the present invention has been described in detail above. For those skilled in the art, there will be changes in the specific implementation and application scope based on the ideas of the embodiments of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for identifying abnormally hypermethylated regions in cancer methylation data, comprising: Obtain methylation sequencing data from normal and cancerous tissues; Abnormally hypermethylated regions were screened from the methylation sequencing data of the normal and cancer tissues based on reference UMRs. The reference UMRs were identified using the following method: Step 1: For each WGBS matrix, use BSMAP to trim the transition, low-quality and repetitive sequences with the default threshold, and align the sodium bisulfite-treated reads to the human genome; use the coverage of 4 reads to ensure the accuracy of CpG methylation detection; calculate the methylation rate of each CPG covered by at least 4 reads using the percentage value module in BSMAP. Step 2: UMRs were determined by at least four consecutive hypomethylated CpGs. To reduce the impact of sparse CPG density on UMR detection in the HMM-based model, UMRs with Obs / Exp values ​​< 0.1 were removed. Hypomethylated CpGs were determined by a hypo-average methylation ratio of 10%. Step 3: By merging overlapping UMRs from multiple samples, redundant UMRSs are reduced from multiple tissue and tumor WGBS profiles. To describe the enrichment distribution of genome-wide UMRs in tissue and tumor samples, the UMR frequency (UOF) of crossover fragments in N samples is: UMR percentage score represents the percentage of UMR in population-sized samples of normal tissue and tumor. The higher the UOF, the more conservative the UMR in the population-sized sample. Conversely, a decrease in UOF represents a shortening or loss of UMR, which indicates that these regions are hypermethylated at the population scale. Step 4: Based on the Poisson test, detect reference UMRs from the UOF spectrum of the sample population. These reference UMRs were identified from normal tissues and tumors, respectively. Calculating the proportion of hypermethylated CpG sites in the anomalously hypermethylated regions relative to CpG sites in the reference UMRs specifically includes: The hypermethylation state within the abnormally hypermethylated region is determined using machine learning algorithms and / or statistical methods; the CpG site within the hypermethylation state is the hypermethylated CpG site; the methylation state also includes: hypomethylated, no difference; The CpG sites in the Hyper methylation state are merged into the methylation region to obtain the methylation region after merging CpG sites, which is the abnormally hypermethylated region. Based on the bimodal distribution curve of the proportion of hypermethylated CpG sites in the reference UMRs, the proportion of hypermethylated CpG sites in the methylated region after merging CpG sites is calculated. Based on the aforementioned ratio, the abnormally hypermethylated regions are divided into phUMRs or fhUMRs; phUMRs indicate that CpG hypermethylation occurs only in a portion of the UMRs, while fhUMRs indicate that hypermethylation occurs throughout the entire UMRs.

2. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 1, characterized in that, When using machine learning algorithms to determine the methylation state, the method also includes modeling the emission probability matrix of the methylation state using mathematical distribution methods, calculating the mean and variance based on the methylation differences between normal tissue and cancer tissue methylation sequencing data, and calculating the methylation levels of adjacent CpG sites to obtain the transition probability.

3. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 1, characterized in that, When determining the methylation status using statistical methods, a methylation matrix for each CpG site in each of the reference UMRs is established; the p-value for each CpG site in the reference UMRs is calculated using a two-tailed test, and the state of CpG sites with an FDR-corrected p-value less than 0.05 and an absolute methylation difference greater than a first threshold is defined as Hyper, with the first threshold being 0.

2.

4. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 2, characterized in that, The machine learning algorithm includes one or more of the following: Hidden Markov Model, Hotspot Augmentation Algorithm, Sliding Window Algorithm, preferably Hidden Markov Model.

5. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 2, characterized in that, The mathematical distribution method includes one or more of the following: Gaussian distribution, beta distribution, binomial distribution, preferably Gaussian distribution.

6. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 1, characterized in that, The CpG sites within the Hypermethylation state include: CpG sites within the Hypermethylation state determined using a machine learning algorithm; or, CpG sites within the Hypermethylation state determined using a statistical method; or, the CpG sites obtained by taking the union or intersection of the CpG sites within the Hypermethylation state determined using a machine learning algorithm and the CpG sites within the Hypermethylation state determined using a statistical method.

7. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 1, characterized in that, The methylation sequencing data were WGBS data.

8. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 1, characterized in that, The step of classifying the abnormally hypermethylated regions into phUMRs or fhUMRs based on the ratio includes: the ratio being less than or equal to or less than a second threshold as phUMRs, and the ratio being greater than or equal to or greater than the second threshold as fhUMRs, wherein the second threshold is 0.

8.

9. The method for identifying abnormally hypermethylated regions in cancer methylation data according to any one of claims 1-8, characterized in that, The results of the phUMR typing were used to determine the activation of oncogenes in the sample.

10. The method for identifying abnormally hypermethylated regions in cancer methylation data according to any one of claims 1-8, characterized in that, Based on the typing results of the phUMRs, one or more of the following results can be obtained in the sample: the gene in the phUMRs on the promoter is upregulated, the upregulated gene is rich in oncogenes, the upregulated gene on the promoter has increased active histone modification signal and decreased repressive histone signal.

11. The method for identifying abnormally hypermethylated regions in cancer methylation data according to any one of claims 1-8, characterized in that, The method further includes defining the hypermethylated regions within the phUMRs as partially hyper, and the hypomethylated regions within the phUMRs as flanying UMRs, compared to the reference UMRs.

12. The method for identifying abnormally hypermethylated regions in cancer methylation data according to claim 11, characterized in that, Based on the results of the partially Hyper method, significant enrichment of H3K4me3 and H3K27ac was observed, along with a decrease in the activity modification signals of related phUMRs associated with promoter upregulation and downregulation; based on the results of the flanking UMR method, an increase in the activity modification signals of phUMRs associated with promoter upregulation was observed.

13. A device for identifying abnormally hypermethylated regions in cancer methylation data, the device comprising: Memory and processor; The memory is used to store program instructions; The processor is used to invoke program instructions, which, when executed, are used to perform the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data according to any one of claims 1-12.

14. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data as described in any one of claims 1-12.

15. A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method for identifying abnormally hypermethylated regions in cancer methylation data according to any one of claims 1-12.

Images