Method and equipment for determining genome imprinting sample regulation region

Abstract

The invention provides a method and equipment for determining a genome imprinting sample regulation region. The method comprises the following steps: acquiring methylation sequencing data of an individual, which is composed of a plurality of sequencing reads; dividing the genome reference sequence of the individual into a plurality of primary windows; selecting a primary imprint-like regulatory region from the plurality of primary windows based on the methylation level of the CpG site in the primary windows; and extending the primary imprint-like regulatory region, where the primary imprint-like regulatory region is a primary window that satisfies the following criteria: (a) sequencing reads that fall into the primary window can be classified into a high methylation group and a low methylation group, the methylation level of the high methylation group being higher than the methylation level of the low methylation group; and (b) the difference in the number of sequencing reads between the hypermethylation group and the hypomethylation group does not exceed a predetermined threshold. A genome imprinting sample regulation region can be effectively determined by selecting a nucleic acid sequence meeting specific conditions.

Description

Technical Field

[0001] The present application relates to the field of biotechnology, specifically to the field of genomics, and more specifically to a method, device, server and computer-readable storage medium for determining a genomic imprinting-like regulatory region. Background Art

[0002] Genomic imprinting is an important phenomenon in mammalian growth and development, which refers to the specific expression of paternal or maternal alleles. In almost all known cases in mammals, the expression of such imprinted genes has been determined to be regulated by allele-specific methylation (ASM) of certain cis-acting elements. Therefore, the identification of imprinted genes and their regulatory mechanisms is very important for understanding mammalian development and abnormal genomic imprinting diseases.

[0003] High-throughput sequencing technology has made it possible to analyze imprinted genes on a large scale. Whole-genome bisulfite sequencing (WGBS) is the most useful method to study the DNA methylome. In the past decade, by using WGBS data and its associated heterozygous SNP information, researchers have detected many allele-specific methylation regions (AMRs) in the human and mouse genomes.

[0004] Although these studies provide valuable resources for our understanding of genomic imprinting mechanisms and allele expression mechanisms in mammalian cells, there are two major practical limitations of the current popular methods: First, studies do not always obtain sample-specific heterozygous SNP information, and in some specific studies, it is possible that some AMRs do not naturally contain heterozygous SNP information. Second, using only sample-specific methylation data and extremely relevant heterozygous SNP information cannot distinguish the origin of AMR (including parental imprinting, germline sequence dependence, and random generation).

[0005] Recently, some AMR discovery methods that do not rely on SNPs have been published. For example, Amrfinder (F. Fang, E. Hodges, A. Molaro, M. Dean, GJ Hannon, A. Smith, Genomic landscape of human allele-specific DNA methylation, Proc Natl Acad Sci USA 109 (2012) 7332-7337.10.1073 / pnas.1201310109.) contains two statistical models (one model assumes that both alleles are methylated, and the other assumes that the two alleles have different methylation states) to evaluate whether the region is ASM by comparing the likelihood of the two models. However, this method is based on the assumption that the methylation levels of single or paired CpG sites in the genome are independent of each other, so a large number of unknown variables will increase the complexity and time consumption of the calculation, and most importantly, it will increase the uncertainty of AMR identification; and this type of method cannot distinguish the origin of AMR.

[0006] Therefore, current research on genomic imprinting still needs to be improved. Summary of the invention

[0007] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, one object of the present invention is to provide a method for effectively determining genomic imprinting-like regulatory regions.

[0008] In a first aspect of the present application, a method for determining a genomic imprinting-like regulatory region is proposed. According to an embodiment of the present application, the method includes: obtaining methylation sequencing data of an individual, wherein the methylation sequencing data is composed of a plurality of sequencing reads; dividing a genomic reference sequence of a species corresponding to the individual into a plurality of primary windows, wherein the primary windows contain CpG sites; selecting a primary imprinting-like regulatory region from the plurality of primary windows based on the methylation status of the CpG sites in the primary windows; and optionally, extending the primary imprinting-like regulatory region with adjacent primary windows to obtain the genomic imprinting-like regulatory region, wherein the primary imprinting-like regulatory region is the primary window that meets the following criteria: (a) the sequencing reads falling into the primary window can be classified into a high methylation group and a low methylation group, and the methylation level of the high methylation group is higher than the methylation level of the low methylation group; and (b) the difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed a predetermined threshold.

[0009] According to an embodiment of the present application, a genomic imprinting-like regulatory region can be effectively determined by selecting a nucleic acid sequence that satisfies a specific condition. Specifically, the inventors of the present application have found that in a single cell population, the sequencing reads (reads) obtained by methylation sequencing, such as the sequencing reads obtained by whole genome bisulfite sequencing (WGBS), except for the allele-specific methylation (ASM) phenomenon, CpG methylation levels are usually consistent within a certain genome interval. Thus, according to an embodiment of the present application, if the following conditions are met, a genomic region of a certain length can be used as a candidate region for a genomic imprinting-like regulatory region: (i) the methylation sequencing reads falling into a given genomic region can be divided into two different groups (a high methylation group and a low methylation group), and (ii) the sequencing reads in the high methylation group and the low methylation group are similar. Thus, a genomic imprinting-like regulatory region can be effectively determined by the method according to an embodiment of the present application.

[0010] According to an embodiment of the present application, the above method for determining a genomic imprinting-like regulatory region may also have at least one of the following additional technical features:

[0011] According to an embodiment of the present application, the methylation sequencing data is single-base resolution DNA methylation sequencing data.

[0012] According to an embodiment of the present application, in (a), the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.5, preferably not less than 0.7, more preferably not less than 0.8, for example not less than 0.85 or not less than 0.90.

[0013] According to an embodiment of the present application, in (b), the ratio of the number of sequencing reads with fewer sequencing reads in the hypermethylation group and the hypomethylation group to the total number of sequencing reads falling into the primary window is not less than 20%, preferably not less than 25%, more preferably not less than 30%, for example not less than 40%, or

[0014] The difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed 10%, preferably does not exceed 5%.

[0015] According to an embodiment of the present application, the primary imprinting regulatory region and the adjacent primary window are extended according to the following criteria to obtain the genomic imprinting-like regulatory region:

[0016] (i) the distance between the primary imprinting-like regulatory region and the adjacent primary window is no more than 1000 bp, preferably no more than 800 bp, more preferably no more than 700 bp, and most preferably no more than 600 bp;

[0017] (ii) the primary imprinting-like regulatory region and each of the adjacent primary windows contain at least 5 CpG sites, preferably at least 10 CpG sites;

[0018] (iii) after extending the primary imprinted regulatory region and the adjacent primary window, the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.8, such as not less than 0.85 or not less than 0.90; and

[0019] (iv) After extending the primary imprinting regulatory region and the adjacent primary window, the number of sequencing reads in the high methylation group and the low methylation group with fewer sequencing reads accounts for no less than 30% of the total number of sequencing reads falling into the primary window, for example, no less than 40%.

[0020] According to an embodiment of the present application, for a given primary window, the methylation levels of the high methylation group and the low methylation group are determined by performing a maximum expectation algorithm on the sequencing reads falling within the primary window.

[0021] According to an embodiment of the present application, the maximum expectation algorithm includes: assigning initialization values ​​to the methylation levels of the high methylation group A and the low methylation group B, respectively; and using the methylation levels of at least a portion of the sequencing reads that fall into the primary window, performing at least one iterative calculation update on the methylation levels of the high methylation group A and the methylation levels of the low methylation group, so as to obtain the final methylation levels of the high methylation group and the low methylation group of the given primary window.

[0022] According to an embodiment of the present application, the maximum expectation algorithm includes: the maximum expectation algorithm includes: (1) assigning initialization values ​​to the methylation level of the high methylation group A and the methylation level of the low methylation group B respectively; (2) for each sequencing read falling into the primary window, based on the methylation level of the sequencing read, calculating the probability of the sequencing read being classified into the high methylation group A and the low methylation group B, so as to group the sequencing reads falling into the primary window between the high methylation group A and the low methylation group; (3) based on the grouping result obtained in step (2), determining the methylation level of the high methylation group A and the low methylation group B respectively; (4) returning the methylation level obtained in step (3) to step (1) to update the initialization value, and repeating steps (1) to (3) at least once, so as to obtain the final methylation level of the high methylation group and the low methylation group of the given primary window.

[0023] Specifically, according to an embodiment of the present application, the maximum expectation algorithm may include: first, randomly assigning an initialization value (such as 0.1, 0.9) to the methylation level of the high methylation group A and the methylation level of the low methylation group B; then, for each sequencing read falling into the primary window, according to the methylation level of the sequencing read and the methylation levels of the high methylation group A and the low methylation group B, respectively calculate the probability P1 that the read comes from group A and the probability P2 that the read comes from group B. According to the principle of maximum likelihood estimation, it is believed that the hypothesis with the highest probability always has the highest probability of occurring, that is, the reads are divided into to the group with the highest probability among P1 and P2; after each sequencing read is divided into one of the AB groups, the methylation level of each group is recalculated separately according to all the reads in the AB groups; finally, based on the methylation levels of the high methylation group A and the low methylation group B just calculated, each sequencing read falling into the primary window is iteratively calculated, so as to determine the sequencing reads respectively contained in the high methylation group and the low methylation group of the given primary window, and then obtain the final methylation levels of the high methylation group and the low methylation group of the given primary window.

[0024] According to an embodiment of the present application, the iterative calculation is performed at least 100 times, or the iterative calculation is stopped when the difference between the methylation level of the previous high methylation group and the methylation level of the low methylation group is less than 0.0001 at the same time.

[0025] According to an embodiment of the present application, it further includes: excluding tissue-specific DNA methylation pattern regions from the genomic imprinting-like regulatory regions to obtain the screened genomic imprinting-like regulatory regions; and judging the degree of similarity between the screened genomic imprinting-like regulatory regions and known imprinting regulatory regions, so as to output the final genomic imprinting-like regulatory regions.

[0026] According to an embodiment of the present application, the similarity determination is performed based on the following formula:

[0027] p=1 / (1+e-(-31.33+20.2256*diff_score+0.0043*length+13.2152*ratio+0.0903*cg)),

[0028] in,

[0029] p represents the similarity between the screened genomic imprinted regulatory region and the known imprinted regulatory region;

[0030] diff_score represents the absolute value (diff_score) of the difference in methylation levels between the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region;

[0031] cg represents the number of CpG sites in the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region;

[0032] ratio represents the ratio of the number of sequencing reads of the high methylation group to the low methylation group in the screened genomic imprinting-like regulatory region; and

[0033] Length represents the length of the candidate region in the screened genomic imprinting-like regulatory region.

[0034] According to an embodiment of the present application, the length of the primary window is 200 to 500 bp, preferably 300 bp.

[0035] In a second aspect of the present application, the present application proposes a device for determining a genomic imprinting-like regulatory region. According to an embodiment of the present application, the device includes: an acquisition module for acquiring methylation sequencing data of an individual, wherein the methylation sequencing data is composed of a plurality of sequencing reads; a primary window division module for dividing the genomic reference sequence of the species corresponding to the sample into a plurality of primary windows, wherein the primary windows contain CpG sites; a primary imprinting-like regulatory region determination module for selecting a primary imprinting-like regulatory region from the plurality of primary windows based on the methylation status of the CpG sites in the primary windows; and a genomic imprinting-like regulatory region determination module for extending the primary imprinting-like regulatory region with the adjacent primary windows to obtain the genomic imprinting-like regulatory region, wherein the primary imprinting-like regulatory region is the primary window that meets the following criteria: (a) the sequencing reads falling into the primary window can be classified into a high methylation group and a low methylation group, wherein the methylation level of the high methylation group is higher than the methylation level of the low methylation group; and (b) the difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed a predetermined threshold.

[0036] By using this device, the method for determining genomic imprinting-like regulatory regions described in the first aspect above can be effectively implemented, thereby effectively determining genomic imprinting-like regulatory regions by selecting nucleic acid sequences that meet specific conditions.

[0037] According to an embodiment of the present application, the above-mentioned device for determining a genomic imprinting-like regulatory region may also have the following additional technical features:

[0038] According to an embodiment of the present application, the genomic imprinting-like regulatory region determination module extends the primary imprinting regulatory region and the adjacent primary window according to the following criteria to obtain the genomic imprinting-like regulatory region:

[0039] (i) the distance between the primary imprinting-like regulatory region and the adjacent primary window is no more than 1000 bp, preferably no more than 800 bp, more preferably no more than 700 bp, and most preferably no more than 600 bp;

[0040] (ii) the primary imprinting-like regulatory region and each of the adjacent primary windows contain at least 5 CpG sites, preferably at least 10 CpG sites;

[0041] (iii) after extending the primary imprinted regulatory region and the adjacent primary window, the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.8, such as not less than 0.85 or not less than 0.90; and

[0042] (iv) After extending the primary imprinting regulatory region and the adjacent primary window, the number of sequencing reads in the high methylation group and the low methylation group with fewer sequencing reads accounts for no less than 30% of the total number of sequencing reads falling into the primary window, for example, no less than 40%.

[0043] According to an embodiment of the present application, it further includes: a screening module, used to exclude tissue-specific DNA methylation pattern regions from the genomic imprinting-like regulatory region to obtain the screened genomic imprinting-like regulatory region; and a similarity judgment module, used to judge the similarity between the screened genomic imprinting-like regulatory region and the known imprinting regulatory region, so as to output the final genomic imprinting-like regulatory region.

[0044] According to an embodiment of the present application, the similarity determination is performed based on the following formula:

[0045] p=1 / (1+e-(-31.33+20.2256*diff_score+0.0043*length+13.2152*ratio+0.0903*cg)),

[0046] in,

[0047] p represents the similarity between the screened genomic imprinted regulatory region and the known imprinted regulatory region;

[0048] diff_score represents the absolute value (diff_score) of the difference in methylation levels between the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region;

[0049] cg represents the number of CpG sites in the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region;

[0050] ratio represents the ratio of the number of sequencing reads of the high methylation group to the low methylation group in the screened genomic imprinting-like regulatory region; and

[0051] Length represents the length of the candidate region in the screened genomic imprinting-like regulatory region.

[0052] In the third aspect of the present application, the present application proposes a server, which includes a processor and a memory, and the memory stores a computer program. When the computer program is executed by the processor, the method for determining genomic imprinting-like regulatory regions described in the first aspect is implemented.

[0053] In the fourth aspect of the present application, the present application proposes a computer-readable storage medium containing a computer program, characterized in that when the computer program is executed by one or more processors, the method for determining genomic imprinting-like regulatory regions described in the first aspect is implemented.

[0054] According to an embodiment of the present application, the use of an AMR recognition algorithm that does not rely on SNP (also referred to herein as an AIMER algorithm) is based on the discovery that DNA methyltransferases always work in specific regions rather than single CpG sites. Therefore, the inventors of the present application have proposed a criterion for judging genomic imprinting-like regulatory regions, that is, except for the allele-specific methylation (ASM) phenomenon, CpG methylation levels are usually consistent within a certain genomic interval. Therefore, according to an embodiment of the present application, a genomic region of a certain length can be used as a candidate region for a genomic imprinting-like regulatory region if the following conditions are met: (i) the methylation sequencing reads falling into a given genomic region can be divided into two different groups (a high methylation group and a low methylation group), and (ii) the number of sequencing reads in the high methylation group and the low methylation group is similar. According to the embodiments of the present application, by using the F1 i methylation group data generated by sequencing by Ren Bing's research group as input (W. Xie, CL Barr, A. Kim, F. Yue, AY Lee, J. Eubanks, E L Dempster, B. Ren, Base-resolution analyses of sequence and parent-of-origin dependent DNA methylation in the mouse genome, Cell 148 (2012) 816-831.10.1016 / j. cell.2011.12.035.), it is proved that compared with other methods, the AIMER algorithm according to the embodiments of the present application performs very well in detecting known imprinted regulatory regions. In addition, according to the embodiments of the present application, the inventors also found that the AMR identified by the AIMER algorithm can be distinguished from the sequence-dependent AMR identified by Ren Bing's research group (i.e., the strain-specific AMR that appears due to different genetic backgrounds), which shows that the means of the present application can be effectively used to find parental origin AMR. In other words, the means according to the embodiments of the present application can be effectively used to obtain genomic imprinted regulatory regions in the genome.

[0055] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0057] Figure 1 It is a flowchart of a method for determining a genomic imprinting-like regulatory region according to one embodiment of the present application;

[0058] Figure 2 is a flow chart of a method for determining a genomic imprinting-like regulatory region according to another embodiment of the present application;

[0059] Figure 3 is a schematic diagram of the structure of a device for determining a genomic imprinting-like regulatory region according to an embodiment of the present application;

[0060] Figure 4 is a schematic structural diagram of a device for determining a genomic imprinting-like regulatory region according to another embodiment of the present application;

[0061] Figure 5 is a flow chart of a method for determining a genomic imprinting-like regulatory region according to another embodiment of the present application; and

[0062] Figure 6 It is a schematic diagram of a process of extending a primary imprinting-like regulatory region to obtain a genomic imprinting-like regulatory region according to an embodiment of the present application.

[0063] Figure 7 A comparison between the algorithm AIMER according to an embodiment of the present application and other methods is shown. Figure 7 A: ROC curve, the larger the area under the curve, the better the performance of the method; Figure 7 B and Figure 7 In C: mCG / CG is the average methylation level of the region, AMR is whether the region is identified as an AMR region by different methods, and Score is the statistical results of different methods (whether the region is AMR is further inferred based on these statistical results).

[0064] Figure 8 A comparison of the region- and sequence-dependent AMRs found by the algorithm AIMER according to an embodiment of the present application is shown. Figure 8 A: Comparison of the two box plots. Each point is the average number of CGs per 300 bp in a region. Figure 8 B: The x-axis is the ratio of the number of sequencing reads contained in the two chromatids (high methylation group and low methylation group) corresponding to each region calculated by AIMER (the group with smaller number is the numerator and the group with larger number is the denominator), and the y-axis is the diff_score calculated by AIMER; Figure 8 C: Principal component analysis (PCA) of AMRs obtained by different methods, where the red Known is the ICR that has been reported and supported by experiments, the blue Novel is the ICR newly discovered by Ren Bing's research group, the purple AIMER is the ICR found by the inventor's software, and Sequence-dependent is the strain-specific AMR found by Ren Bing's research group through experiments; Figure 8D: mCG / CG is the average methylation level of the region, AMR is whether the region is identified as an AMR region by different methods, and Score is the statistical result of different methods (whether the region is AMR is further inferred based on these statistical results). DETAILED DESCRIPTION

[0065] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0066] definition

[0067] Unless otherwise specified in this article, the term "genomic imprinting" is used interchangeably with the term "genetic imprinting" or "imprinting regulatory region", which refers to the heritable epigenetic modification phenomenon in which a specific gene or gene cluster in an offspring individual only expresses the allele from one side of the paternal or maternal source, while the parental allele on the other side is silenced. Generally speaking, in general diploid organisms such as humans, there are two copies of the genome in the somatic cells. Usually, alleles in these two copies of the genome can be expressed, but a small number (less than 1%) of genes will be affected by imprinting, making one of the alleles ineffective. For example, an insulin-like growth factor IGF-2 gene, only the allele from the father can be expressed. Imprinting means that the expression of a certain allele is suppressed. If the imprinting state of an imprinted gene changes, it will cause the co-expression or simultaneous silencing of both alleles, affecting life activities. Therefore, since imprinted genes have been widely discovered in higher animals, more and more studies have been conducted in the fields related to development and genetic diseases. The inventors of the present invention believe that these genomic imprinting phenomena are related to differential DNA methylation. In other words, this methylation imprinting state is established on only one gamete and maintained on only one parental chromosome of the embryo. In addition, the establishment and maintenance of differential DNA methylation may be regulated by DNA methyltransferase (Dnmt).

[0068] Unless otherwise specified herein, the term "genomic imprinting-like regulatory region" refers to a region in the genome whose regulation is the same or similar to that of "genomic imprinting", that is, the region only expresses the allele from the paternal or maternal side, while the parental allele on the other side is silenced. By searching for such genomic imprinting-like regulatory regions in the genome, a powerful tool can be provided for subsequent research on the occurrence, diagnosis and treatment of diseases (such as cancer, development-related diseases such as Beckwith-Wiedemann syndrome and Silver-Russells syndrome, etc.).

[0069] Unless otherwise specified in this article, the term "CpG site" is used interchangeably with the term "CG site" and refers to a region of DNA where the base sequence appears as cytosine followed by guanine, where "CpG" is the abbreviation of "-C-phospho-G-", which means that the phosphodiester bond connects cytosine and guanine, with C at the 5' end and G at the 3' end. Among them, the cytosine in the CpG site can be methylated to 5-methylcytosine. In individuals such as mammals, methylation of CpG sites within genes will change the expression of this gene, and the study of this expression regulation is an important part of epigenetics. The enzyme involved in adding methyl groups is called DNA methyltransferase.

[0070] Unless otherwise specified in this article, the term "methylation sequencing" refers to a sequencing method that can distinguish whether cytosine in a nucleic acid molecule is methylated. For example, whole genome methylation sequencing (WGBS) mainly converts unmethylated cytosine in the genome into uracil through bisulfite, and then combines it with a high-throughput sequencing platform to measure the methylation level of cytosine in the entire genome and accurately analyze the methylation status of each cytosine.

[0071] Unless otherwise specified in this article, the term "sequencing read" is used interchangeably with the term "read" or "read segment", which refers to a nucleic acid sequence obtained during sequencing. This sequence is referred to as a "sequencing read segment" or "read" in this article. Usually, the read length of the sequencing reads of first-generation sequencing and third-generation sequencing is between several thousand and tens of thousands of bp, and the sequencing reads of second-generation sequencing are relatively shorter, averaging tens to hundreds of bp.

[0072] Unless otherwise specified in this article, the term "genomic reference sequence" refers to the genome sequence of the species corresponding to the known sample. It can be a genome sequence obtained through public channels or obtained through sequencing assembly. It can be the entire sequence of the genome or a partial sequence of the genome of interest. For example, when analyzing human samples, multiple versions of human genome sequences provided by public databases can be used, such as hg19.

[0073] Unless otherwise specified herein, the term "window" refers to a continuous sequence on a genomic reference sequence.

[0074] Unless otherwise specified herein, the expression "sequencing reads falling into the window" refers to sequencing reads that can be aligned with the nucleic acid sequence within the window, and such alignment can be performed using conventional alignment software. Sometimes sequencing reads may be aligned with multiple positions on the genome, but sometimes, as needed, for the convenience of analysis, only sequencing reads that can be aligned to only one position can be selected, and such reads are also called "uniquely aligned sequencing reads".

[0075] Unless otherwise specified herein, the term "methylation level" is a measure used to characterize the amount of methylation at CpG sites in a certain region (e.g., a sequencing read or window). For example, it can be expressed as the percentage of methylated CpG sites to all CpG sites in the region. For example, in a given region, which contains 100 CpG sites, 40 of which are methylated, the methylation level of the region is 40% or 0.4.

[0076] The technical solution of the present application is described in detail below with reference to the accompanying drawings.

[0077] In the first aspect of the present invention, the present application proposes a method for determining a genomic imprinting-like regulatory region, referring to Figure 1 and 2 , Figure 5 and 6 According to an embodiment of the present application, the method includes:

[0078] S100: Acquire individual methylation sequencing data, which consists of multiple sequencing reads

[0079] In this step, there is no particular restriction on the source of methylation sequencing data, and public methylation sequencing data that has been sequenced before can be used, or sequencing data obtained by independently performing methylation sequencing can be used. In addition, there is no particular restriction on the sequencing platform, and it can be any common first-generation, second-generation or third-generation sequencing platform. Those skilled in the art can mainly convert the cytosine that does not undergo methylation in the genome into uracil by bisulfite (Bisulfite), and then combine the high-throughput sequencing platform to measure the methylation level of cytosine within the whole genome, and accurately analyze the methylation state of each cytosine. According to one embodiment of the present application, methylation sequencing data can be used as DNA methylation sequencing data with a single base resolution. Thus, the accuracy of subsequent analysis is further improved, and the efficiency of obtaining genomic imprinting-like regulatory regions is improved.

[0080] S200: Divide the individual's genomic reference sequence into multiple primary windows, each containing CpG sites

[0081] In this step, at least a portion of the reference genome sequence of the species corresponding to the methylation sequencing data is sequenced to obtain a plurality of primary windows. According to an embodiment of the present application, the length of the primary window is 200 to 500 bp, preferably 300 bp. For subsequent analysis of the methylation level of CpG, these primary windows contain CpG sites, and preferably the number of CpG sites exceeds a predetermined threshold, for example, not less than 5, preferably not less than 10 CpG sites in the primary window.

[0082] According to an embodiment of the present application, the whole genome reference sequence can be divided into a plurality of primary windows of 200 to 500 bp in length containing CpG sites by a sliding window method. After obtaining these primary windows, the primary windows are filtered to remove the primary windows whose CG number is lower than a predetermined threshold, for example, only primary windows containing more than 10 CpG sites are selected. Of course, it can be understood by those skilled in the art that those skilled in the art can determine the sequencing reads falling into the primary window by conventional comparison means after comparing the sequencing data with the reference sequence.

[0083] S300: Selecting primary imprinted regulatory regions from multiple primary windows based on the methylation status of CpG sites in the primary windows

[0084] According to an embodiment of the present application, after the reference sequence is divided into a plurality of primary windows, the sequencing data may be compared with the reference sequence to obtain sequencing reads falling into the primary windows.

[0085] According to the embodiments of the present application, the inventors of the present application believe that DNA methyltransferase always works in a specific region rather than a single CpG site, and thus, the inventors of the present application propose a criterion for judging the whole region of genomic imprinting-like regulation, that is, except for the allele-specific methylation (ASM) phenomenon, the CpG methylation level is usually consistent within a certain genomic interval. Thus, according to the embodiments of the present application, a genomic region of a certain length can be used as a candidate region for a genomic imprinting-like regulatory region if the following conditions are met: (i) the methylation sequencing reads falling into a given genomic region can be divided into two different groups (a high methylation group and a low methylation group), and (ii) the number of sequencing reads in the high methylation group and the low methylation group is similar.

[0086] According to an embodiment of the present application, the primary imprinted regulatory region is a primary window that meets the following criteria: (a) the sequencing reads falling into the primary window can be classified into a high methylation group and a low methylation group, and the methylation level of the high methylation group is higher than the methylation level of the low methylation group; and (b) the difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed a predetermined threshold. According to an embodiment of the present application, in (a), the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.5, preferably not less than 0.7, more preferably not less than 0.8, for example, not less than 0.85 or not less than 0.90. According to an embodiment of the present application, in (b), the number of sequencing reads with fewer sequencing reads in the high methylation group and the low methylation group accounts for a proportion of the total number of sequencing reads falling into the primary window of not less than 20%, preferably not less than 25%, more preferably not less than 30%, for example not less than 40%, or the difference in the number of sequencing reads in the high methylation group and the low methylation group does not exceed 10%, preferably not more than 5%.

[0087] According to an embodiment of the present application, for a given primary window, the methylation levels of the high methylation group and the low methylation group are determined by performing a maximum expectation algorithm on the sequencing reads falling into the primary window. According to an embodiment of the present application, the maximum expectation algorithm includes: using the initialization of the high methylation group methylation level and the initialization of the low methylation group methylation level, the methylation levels of the high methylation group and the low methylation group are iteratively calculated to obtain the final methylation levels of the high methylation group and the low methylation group of the given primary window. According to an embodiment of the present application, the iterative calculation is performed at least 100 times, or the iterative calculation is stopped when at least one of the differences before and after the iteration of the methylation level of the high methylation group and the methylation level of the low methylation group is less than 0.0001. Preferably, the iterative calculation is stopped when the differences before and after the iteration of the methylation level of the high methylation group and the methylation level of the low methylation group are both less than 0.0001.

[0088] Specifically, according to an embodiment of the present application, for a given initial window and sequencing reads falling into the window, an expectation maximization and maximum likelihood estimation algorithm model can be used to divide the sequencing reads of a single segment into a high methylation group A and a low methylation group B.

[0089] More specifically, according to the embodiment model of the present application, it is first assumed that the methylation level of the initialized high methylation group A is m1, and the methylation level of the low methylation group B is m2. When the first read segment R1 is read, the number of methylated CG sites in the current read segment R1 is M1, and the number of unmethylated CG sites is U1. When the current read segment R1 is calculated to come from the high methylation group A according to the binomial distribution, the probability of observing the current read segment R1 is P1 = (R1|A) = m1 M1 .(1-m1) U1Similarly, when the current read segment R1 comes from the low-methylation group B, the probability of observing the current read segment R1 is P2 = (R1|B) = m2 M1 .(1-m2) U1 .

[0090] When read segment R1 is observed, the probability that the current read segment R1 comes from group A can be calculated according to Bayes' theorem. Similarly, the probability that the current read segment R1 comes from group B is Where α is the probability that the current read segment R1 comes from the high methylation group A, and 1-α is the probability that the current read segment R1 comes from the low methylation group B.

[0091] Assuming that the current read segment R1 comes from the highly methylated group A, the expected value of the number of methylated CG sites is E AM(1) =Q1.M1, the expected number of unmethylated CG sites is E AU(1) =Q1.U1; Similarly, assuming that the current read segment R1 comes from the low-methylation group B, the expected value of the number of methylated CG sites is E BM(1) =Q2.M1, the expected number of unmethylated CG sites is E BU(1) =Q2.U1.

[0092] After the calculation of the first sequencing read is completed, the remaining reads in the region (the 2nd to the nth reads) are calculated to complete the first iteration. AM(n) and E BM(n) After that, all the reads in this segment come from the high methylation group A and the expected sum of the number of methylated CG sites is All reads in this segment come from the high methylation group A, and the expected sum of the number of methylated and unmethylated CG sites is Similarly, the expected sum of the number of methylated CG sites in all reads of this segment from the low methylation group B is All reads in this segment come from the low-methylation group B, and the expected sum of the number of methylated and unmethylated CG sites is

[0093] After the first iteration is completed, the methylation level new_m1 of the high methylation group A is calculated, new_m1=nu1 / de1, and the methylation level new_m2 of the low methylation group B is calculated, new_m2=nu2 / de2. The calculated methylation level new_m1 of the high methylation group A replaces the original methylation level m1 of the high methylation group A, and the calculated methylation level new_m2 of the low methylation group B replaces the original methylation level m2 of the low methylation group B. The replaced methylation level m1 of the high methylation group A and the replaced methylation level m2 of the low methylation group B are used for the next iteration calculation, that is, the updated methylation level m1 of the high methylation group A and the methylation level m2 of the low methylation group B are used to continue the above calculation for the read segments 1 to n, so as to obtain the updated methylation level m1 of the high methylation group A and the methylation level m2 of the low methylation group B. When the number of iterations exceeds 100 or the absolute value of the difference between new_m1 and m1 and the absolute value of the difference between new_m2 and m2 are both less than 0.0001, it is considered that the calculated methylation level new_m1 of the high methylation group A and the methylation level new_m2 of the low methylation group B tend to be stable, new_m1 is the final methylation level of the high methylation group A, and new_m2 is the final methylation level of the low methylation group B.

[0094] After calculating the final methylation level new_m1 of the hypermethylated group A and the methylation level new_m2 of the hypomethylated group B in the segment, as described above, the probability value p1 of all reads in the segment coming from the hypermethylated group A and the probability value p2 of all reads coming from the hypomethylated group B are calculated according to the binomial distribution. If p1 is greater than p2, the read is assigned to the hypermethylated group A; if p1 is less than p2, the read is assigned to the hypomethylated group B. The difference in methylation levels between the hypermethylated group A and the hypomethylated group B is the absolute value of the difference between m1 and m2.

[0095] According to an embodiment of the present application, for a given primary window (for example, the CpG content in each window is not less than 5 and the length is greater than 300bp), the methylation level m1 and the methylation level m2 represent the methylation levels in the high methylation group A and the low methylation group B, respectively. In addition, α1 represents the ratio of the number of sequencing reads classified into the high methylation group A to the number of sequencing reads falling into the primary window, and α2 represents the ratio of the number of sequencing reads classified into the low methylation group B to the number of sequencing reads falling into the primary window. α1+α2=1, so for a given primary window, its methylation pattern can be represented by Θ=(m1, m2, α1). It is described above that the final methylation level m1 and methylation level m2 are obtained by iterative expectation maximization algorithm using the initialized methylation level m1 and methylation level m2. According to an embodiment of the present application, the initialized α1 or α2 can also be added as a parameter for iterative update, so as to obtain the final m1, m2, α1, α2 through iterative operation. By using this algorithm, accurate testing can be achieved even at low sequencing depth. For example, the inventors of the present application unexpectedly discovered that very accurate predictions can be achieved at a sequencing depth of 10X, and further increasing the sequencing depth does not contribute much to improving the algorithm.

[0096] S400: Extending the primary imprinting-like regulatory region and the adjacent primary window to obtain the genomic imprinting-like regulatory region

[0097] In this step, after obtaining the primary imprinting-like regulatory region, the primary imprinting-like regulatory region and the adjacent primary window can be extended to obtain the genomic imprinting-like regulatory region. According to an embodiment of the present application, the primary imprinting regulatory region and the adjacent primary window can be extended to obtain the genomic imprinting-like regulatory region according to the following standards:

[0098] (i) the distance between the primary imprinting-like regulatory region and the adjacent primary window is no more than 1000 bp, preferably no more than 800 bp, more preferably no more than 700 bp, and most preferably no more than 600 bp;

[0099] (ii) the primary imprinting-like regulatory region and each of the adjacent primary windows contain at least 5 CpG sites, preferably at least 10 CpG sites;

[0100] (iii) after extending the primary imprinted regulatory region and the adjacent primary window, the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.8, such as not less than 0.85 or not less than 0.90; and

[0101] (iv) After extending the primary imprinting regulatory region and the adjacent primary window, the number of sequencing reads in the high methylation group and the low methylation group with fewer sequencing reads accounts for no less than 30% of the total number of sequencing reads falling into the primary window, for example, no less than 40%.

[0102] According to an embodiment of the present application, extension processing is performed in the following manner, wherein the distance between adjacent primary imprinting-like regulatory regions is called a gap; the first starting candidate window (primary imprinting-like regulatory region), that is, the sequencing reads that fall into the primary window can be classified into a high methylation group and a low methylation group, and the methylation level of the high methylation group is higher than the methylation level of the low methylation group; and the primary window in which the difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed a predetermined threshold. The last primary window from the first starting candidate window to the next gap of more than 600bp is called the end window; the first starting candidate window after the gap of more than 600bp is used as the starting boundary, and the end window is used as the boundary end point. Between the starting boundary and the end point, each starting candidate window is used as the starting point to exhaustively extend all primary windows within the boundary; and the absolute value of the difference in methylation levels between the high methylation group and the low methylation group in the extended region is not less than 0.9; the ratio of the number of sequencing reads between the high methylation group and the low methylation group in the extended region is not less than 0.4; and the genomic imprinting regulatory region contains at least one high GC content primary window, and the high GC content primary window contains at least 10 GC base pairs; if multiple extensions appear in the extended region, only the longest one is left, which is called the candidate region.

[0103] According to an embodiment of the present application, a genomic imprinting-like regulatory region can be effectively determined by selecting a nucleic acid sequence that satisfies a specific condition. Specifically, the inventors of the present application have found that in a single cell population, the sequencing reads (reads) obtained by methylation sequencing, such as the sequencing reads obtained by whole genome bisulfite sequencing (WGBS), except for the allele-specific methylation (ASM) phenomenon, CpG methylation levels are usually consistent within a certain genome interval. Thus, according to an embodiment of the present application, if the following conditions are met, a genomic region of a certain length can be used as a candidate region for a genomic imprinting-like regulatory region: (i) the methylation sequencing reads falling into a given genomic region can be divided into two different groups (a high methylation group and a low methylation group), and (ii) the sequencing reads in the high methylation group and the low methylation group are similar. Thus, a genomic imprinting-like regulatory region can be effectively determined by the method according to an embodiment of the present application.

[0104] According to an embodiment of the present application, it may further include: excluding regions with tissue-specific DNA methylation patterns from genomic imprinted-like regulatory regions to obtain screened genomic imprinted-like regulatory regions; and judging the degree of similarity between the screened genomic imprinted-like regulatory regions and known imprinted regulatory regions, so as to output the final genomic imprinted-like regulatory regions.

[0105] According to an embodiment of the present application, the similarity determination is performed based on the following formula:

[0106] p=1 / (1+e-(-31.33+20.2256*diff_score+0.0043*length+13.2152*ratio+0.0903*cg)),

[0107] in,

[0108] p represents the similarity between the screened genomic imprinted regulatory regions and known imprinted regulatory regions;

[0109] diff_score represents the absolute value of the difference in methylation levels between the high-methylation group and the low-methylation group in the screened genomic imprinting-like regulatory region (diff_score);

[0110] cg represents the number of CpG sites in the high-methylation group and low-methylation group in the screened genomic imprinting-like regulatory regions;

[0111] ratio represents the ratio of the number of sequencing reads in the high methylation group to the low methylation group in the screened genomic imprinting-like regulatory region; and

[0112] Length represents the length of the candidate region in the screened genomic imprinting-like regulatory region.

[0113] According to an embodiment of the present application, Figure 5 is a flow chart of a method for determining a genomic imprinting-like regulatory region according to another embodiment of the present application; Figure 6 It is a schematic diagram of a process of extending a primary imprinting-like regulatory region to obtain a genomic imprinting-like regulatory region according to an embodiment of the present application.

[0114] According to the embodiments of the present application, reference Figure 5 (Figure Same as 1 to 14), the method for determining the genomic imprinting-like regulatory region includes the following modules:

[0115] 1. Represents the input file of the model. The input file is the SAM / BAM file generated after the single-base resolution DNA methylation sequencing data is aligned back to the reference genome;

[0116] 2. Indicates that the sliding window method divides the whole genome reference sequence into a plurality of primary windows of predetermined lengths;

[0117] 3. Indicates the primary window after division;

[0118] 4. Indicates that the EM algorithm is used to group the sequencing reads within the primary window based on the mixed probability model;

[0119] 5. Indicates the sequencing reads after grouping within the primary window;

[0120] 6. The difference in methylation levels between the high methylation group and the low methylation group is calculated as the absolute value of the difference between m1 and m2, diff_score;

[0121] 7. Indicates that starting from the first selected starting candidate window, the primary window with a neighboring distance of no more than 600 bp is extended;

[0122] 8. Indicate candidate regions;

[0123] 9. Indicates that regions with tissue-specific DNA methylation patterns are excluded from candidate regions according to the given list;

[0124] 10. Indicates the similarity between the candidate region excluding tissue specificity and the known imprinted regulatory region;

[0125] 11. Indicates output of the found genomic imprinted regulatory regions;

[0126] 12. Indicates the grouping module;

[0127] 13. Indicates extension module;

[0128] 14. Represents the exclusion module and the similarity calculation module.

[0129] According to the embodiments of the present application, reference Figure 6 , the extension module includes the following steps:

[0130] ① The blue line segment is the primary window; the distance between adjacent primary windows is called a gap; the primary window that meets the absolute value of the methylation level difference greater than 0.85 and the sequencing read segment ratio between the high methylation group and the low methylation group is close to 1:1 is called the starting candidate window;

[0131] ② means starting from the first candidate window after the gap of more than 600bp, and the last primary window before the next gap of more than 600bp after the first candidate window is called the terminal window; starting from the first candidate window, all primary windows between the first candidate window and the terminal window are extended exhaustively to obtain all extended regions;

[0132] ③ means starting from the second starting candidate window, exhaustively extending all primary windows between the second starting candidate window and the end window to obtain all extended regions;

[0133] ④ means starting from the last starting candidate window, exhaustively extending all primary windows between the last starting candidate window and the end window to obtain all extended areas (if the end window is the last starting candidate window, the extended area is still the last starting candidate window);

[0134] ⑤ It means filtering out the areas below the threshold in the extended area. The filtering method is shown in Article 18;

[0135] ⑥ means that for the regions obtained after filtering, the longest region is retained as the candidate region.

[0136] According to the embodiments of the present application, the AMR recognition algorithm (also referred to as AIMER algorithm in this article) that does not rely on SNP is proposed based on the discovery that DNA methyltransferase always plays a role in a specific region rather than a single CpG site. Thus, the inventor of the present application has proposed a judgment criterion for judging the genome imprinting sample regulatory region, that is, except for the allele-specific methylation (ASM) phenomenon, the CpG methylation level is usually consistent within a certain genome interval. Thus, according to the embodiments of the present application, if the following conditions are met, a genome region of a certain length can be used as a candidate region for the genome imprinting sample regulatory region: (i) the methylation sequencing reads falling into a given genome region can be divided into two different groups (a high methylation group and a low methylation group), and (ii) the sequencing reads in the high methylation group and the low methylation group are similar. According to the embodiments of the present application, the F1 i methylation group data generated by the sequencing of the Ren Bing research group is used as input, which proves that compared with other methods, the AIMER algorithm according to the embodiment of the present application performs very well in detecting known imprinting regulatory regions. In addition, according to the embodiments of the present application, the inventors also found that the AMR identified by the AIMER algorithm can be distinguished from the sequence-dependent AMR identified by Ren Bing's research group, which shows that the means of the present application can be effectively used to find parental origin AMR. In other words, the means according to the embodiments of the present application can be effectively used to obtain genomic imprinting-like regulatory regions in the genome.

[0137] In a second aspect of the present application, the present application proposes a device for determining a genomic imprinting-like regulatory region, referring to Figure 3 and Figure 4According to an embodiment of the present application, the device includes: an acquisition module 100, used to acquire methylation sequencing data of an individual, the methylation sequencing data consisting of multiple sequencing reads; a primary window division module 200, used to divide the genomic reference sequence of the species into multiple primary windows, the primary windows containing CpG sites; a primary imprinting-like regulatory region determination module 300, used to select a primary imprinting-like regulatory region from multiple primary windows based on the methylation status of the CpG sites in the primary windows; and a genomic imprinting-like regulatory region determination module 400, used to extend the primary imprinting-like regulatory region to obtain a genomic imprinting-like regulatory region, wherein the primary imprinting-like regulatory region is a primary window that meets the following criteria: (a) the sequencing reads falling into the primary window can be classified into a high methylation group and a low methylation group, and the methylation level of the high methylation group is higher than the methylation level of the low methylation group; and (b) the difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed a predetermined threshold.

[0138] By using this device, the method for determining genomic imprinting-like regulatory regions described in the first aspect above can be effectively implemented, thereby effectively determining genomic imprinting-like regulatory regions by selecting nucleic acid sequences that meet specific conditions.

[0139] According to an embodiment of the present application, the genomic imprinting-like regulatory region determination module is used to extend the primary imprinting regulatory region and the adjacent primary window according to the following criteria to obtain the genomic imprinting-like regulatory region:

[0140] (i) the distance between the primary imprinting-like regulatory region and the adjacent primary window is no more than 1000 bp, preferably no more than 800 bp, more preferably no more than 700 bp, and most preferably no more than 600 bp;

[0141] (ii) the primary imprinting-like regulatory region and each of the adjacent primary windows contain at least 5 CpG sites, preferably at least 10 CpG sites;

[0142] (iii) after extending the primary imprinted regulatory region and the adjacent primary window, the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.8, such as not less than 0.85 or not less than 0.90; and

[0143] (iv) After extending the primary imprinting regulatory region and the adjacent primary window, the number of sequencing reads in the high methylation group and the low methylation group with fewer sequencing reads accounts for no less than 30% of the total number of sequencing reads falling into the primary window, for example, no less than 40%.

[0144] According to an embodiment of the present application, it further includes: a screening module 500, used to exclude tissue-specific DNA methylation pattern regions from the genomic imprinting-like regulatory regions to obtain screened genomic imprinting-like regulatory regions; and a similarity judgment module 600, used to judge the similarity between the screened genomic imprinting-like regulatory regions and the known imprinting regulatory regions, so as to output the final genomic imprinting-like regulatory regions.

[0145] According to an embodiment of the present application, the similarity determination is performed based on the following formula:

[0146] p=1 / (1+e-(-31.33+20.2256*diff_score+0.0043*length+13.2152*ratio+0.0903*cg)),

[0147] in,

[0148] p represents the similarity between the screened genomic imprinted regulatory regions and known imprinted regulatory regions;

[0149] diff_score represents the absolute value of the difference in methylation levels between the high-methylation group and the low-methylation group in the screened genomic imprinting-like regulatory region (diff_score);

[0150] cg represents the number of CpG sites in the high-methylation group and low-methylation group in the screened genomic imprinting-like regulatory regions;

[0151] ratio represents the ratio of the number of sequencing reads in the high methylation group to the low methylation group in the screened genomic imprinting-like regulatory region; and

[0152] Length represents the length of the candidate region in the screened genomic imprinting-like regulatory region.

[0153] In the third aspect of the present application, the present application proposes a server, the server includes a processor and a memory, the memory stores a computer program, when the computer program is executed by the processor, the method of determining the genomic imprinting-like regulatory region of the first aspect is implemented.

[0154] In the fourth aspect of the present application, the present application proposes a computer-readable storage medium containing a computer program, characterized in that when the computer program is executed by one or more processors, the method for determining genomic imprinting-like regulatory regions of the first aspect is implemented.

[0155] It should be noted that the features and technical effects described in this article for different aspects can be used as reference for each other and will not be repeated here.

[0156] Example

[0157] The present application is illustrated below by way of examples, but this should not be construed as limiting the scope of the subject matter of the present application to the following examples. All technologies implemented based on the above content of the present application belong to the scope of the present application. The compounds or reagents used in the following examples can be purchased from commercial sources or prepared by conventional methods known to those skilled in the art; the experimental instruments used can be purchased from commercial sources.

[0158] 1. Dataset

[0159] The methylation sequencing data generated by Ren Bing's research group and stored in GEO with accession number GSE33722 was used as the input for the study of this embodiment. Ren Bing's research group interbred two inbred mouse strains 129X1 / SvJ (129) and Cast / EiJ (Cast), and used the frontal cortex DNA of adult F1 offspring from the initial hybrid 129 (mother) X Cast (father) (denoted as F1i) and the interbred Cast (mother) X 129 (father) (denoted as F1r) to perform WGBS. 1.54 billion (F1i) and 1.33 billion (F1r) unique alignment sequencing reads were generated. In order to distinguish parental source information, 20.4 million SNPs between the 129 genome and the Cast genome were also identified.

[0160] 2. AMR identification

[0161] refer to Figure 5 and 6 , AMR identification was performed according to the algorithm AIMER of the present application, and the F1i methylation group sequencing data was downloaded from GEO. The sequencing data was quality controlled and aligned to obtain the SAM / BAM files. For each window, the diff_score was calculated under the probability model (as described above). Next, for the primary imprinting-like regulatory regions that meet the conditions (diff_score>0.85, the proportion of minor components>0.3), the primary windows with adjacent distances less than 700bps were extended. The extended regions with average diff_score>0.9, number of CG sites>10, and average proportion of minor components after extension>0.4 are called genomic imprinting-like regulatory regions. Finally, genomic imprinting-like regulatory regions that contain only one primary imprinting-like regulatory region are excluded.

[0162] 3. Sequence-dependent AMR identified by Ren Bing’s research group

[0163] S-AS scores and sequence-dependent AMRs were calculated and determined by Ren Bing's research group. Specifically, strain-specific sequencing reads (129 and Cast) pooled from F1i and F1r were used to evaluate each CG by Fisher's exact test to reflect strain-specific allelic DNA methylation biases due to different genetic backgrounds. –log10 (p-value) was considered as the S-AS score. Sequence-dependent AMR refers to CG sites with S-AS scores >= 5, and single CGs are merged into S-AS score regions (with at least 3 or more candidate high S-AS score CGs in adjacent + / -2.5kb windows). The inventors then extracted the top 300 sequence-dependent AMRs (sorted by S-AS scores) and compared them with the AMRs identified by AIMER (using only the F1i methylation group as input, without considering strain information).

[0164] 4. Results

[0165] Identification of AMR by SNP-independent method AIMER

[0166] By using only the F1i methylome as input, AIMER identified 32 AMRs (Table 1). Evaluation was performed in four approaches and the results are summarized in Figure 7 i) AIMER (by F1i methylome, red), ii) P-AS score (by F1i, F1r methylome, as well as parental-SNP information, materials and methods, green), iii)-iv) The primary imprinted regulatory regions were selected in a simple and intuitive way, that is, the average methylation levels of the regions obtained after extending the primary imprinted regulatory regions were in the black range (0.45, 0.55) and blue range (0.4, 0.6), respectively. Known imprinted regulatory regions were considered as correct data sets. The ROC curve shows that AIMER performs very well in discovering ICRs, comparable to the P-AS score ( Figure 7 A). Further analysis showed that AIMER could find almost all known parent-of-origin imprinted AMRs, all of which were at the top of the result list (Supplementary Table 1). However, due to insufficient SNP information, P-AS score selection missed Nnat ( Figure 7 B) In addition, by studying some somatic cell-specific imprinted gene regions (such as Igf2r, Figure 7C), we found that the average methylation level (mCG / CG) was very low or the AIMER diff_score was relatively low, while the P-AS score was high. This may be due to the different input data of the two methods: when calculating the P-AS score, only the methylated sequencing reads containing SNP information were used, and these reads accounted for only 35% of all sequencing reads; while AIMER did not rely on SNP information. All methylated sequencing reads were used. This suggests that it is the data itself, not the calculation method, that limits region identification.

[0167] Table 1:

[0168]

[0169] Comparison with sequence-dependent AMR

[0170] AMR can be caused not only by parental origin (imprinting), but also by sequence dependency. Therefore, the inventors further confirmed whether the novel AMR obtained in this example was caused by imprinting or sequence specificity. The AMR identified by AIMER was compared with the sequence-dependent AMR dataset (generated by Ren Bing's research group, Materials and Methods), and the results are summarized in Figure 8 The results showed that (i) there was no overlap between the AMRs identified by AIMER and the sequence-dependent AMRs ( Figure 8 C). (ii) By calculating the average diff_score and average ratio of sequence-dependent AMRs (generated by AIMER), we found that all these regions retained relatively poor diff_scores or poor ratios (in other words, not close to 1:1) ( Figure 8 B); In addition, the top two sequence-dependent AMRs in the S-AS score ranking have extremely low diff_score ( Figure 8 D). (iii) AIMER-identified AMRs contain relatively more CpG sites than sequence-dependent AMRs ( Figure 8 A). In summary, these results show different characteristics between sequence-dependent AMR and AMR identified by AIMER. Therefore, the results show that the new AMR found by the method AIMER of the present application is more likely to be AMR caused by gene imprinting regulation rather than strain-specific AMR caused by different genetic backgrounds. Or in other words, the method of the present application may be a very valuable option for finding parent-origin AMR.

[0171] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0172] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A method for determining a genomic imprinted regulatory region, characterized in that: include: Acquire methylation sequencing data of an individual, wherein the methylation sequencing data is composed of a plurality of sequencing read segments; Dividing the genome reference sequence of the species corresponding to the individual into a plurality of primary windows, wherein the primary windows contain CpG sites; selecting a primary imprinted-like regulatory region from the plurality of primary windows based on the methylation status of the CpG site in the primary window; and Optionally, the primary imprinting-like regulatory region and the adjacent primary window are extended to obtain the genomic imprinting-like regulatory region. in, The primary imprinted regulatory region is the primary window that meets the following criteria: (a) sequencing reads falling into the primary window can be classified into a high methylation group and a low methylation group, the methylation level of the high methylation group being higher than the methylation level of the low methylation group; and (b) The difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed a predetermined threshold.

2. The method according to claim 1, characterized in that The methylation sequencing data is single-base resolution DNA methylation sequencing data, and the length of the primary window is 200 to 500 bp, preferably 300 bp.

3. The method according to claim 1, characterized in that In (a), the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.5, preferably not less than 0.7, more preferably not less than 0.8, for example not less than 0.85 or not less than 0.

90.

4. The method according to claim 1, characterized in that: In (b), the ratio of the number of sequencing reads with fewer numbers in the high methylation group and the low methylation group to the total number of sequencing reads falling into the primary window is not less than 20%, preferably not less than 25%, more preferably not less than 30%, for example not less than 40%.

5. The method according to claim 1, characterized in that For a given primary window, the methylation levels of the hypermethylated group and the hypomethylated group are determined by performing a maximum expectation algorithm on the sequencing reads falling within the primary window.

6. The method according to claim 5, characterized in that The maximum expectation algorithm includes: Assigning initialization values ​​to the methylation levels of the high methylation group A and the low methylation group B, respectively; and The methylation levels of at least a portion of the sequencing reads falling within the primary window are used to iteratively calculate and update the methylation levels of the high methylation group A and the methylation levels of the low methylation group, so as to obtain the final methylation levels of the high methylation group and the low methylation group of the given primary window.

7. The method according to claim 6, characterized in that The maximum expectation algorithm includes: (1) assigning initialization values ​​to the methylation levels of the high methylation group A and the low methylation group B, respectively; (2) for each sequencing read falling within the primary window, calculating the probability of the sequencing read being classified into the high methylation group A and the low methylation group B based on the methylation level of the sequencing read, so as to group the sequencing reads falling within the primary window between the high methylation group A and the low methylation group; (3) based on the grouping results obtained in step (2), determining the methylation levels of the high methylation group A and the low methylation group B respectively; (4) returning the methylation level obtained in step (3) to step (1) to update the initialization value, and repeating steps (1) to (3) at least once to obtain the final methylation levels of the high methylation group and the low methylation group of the given primary window.

8. The method according to claim 6 or 7, characterized in that: The iterative calculation is performed at least 100 times, or the iterative calculation is stopped when the differences between the methylation level of the high methylation group and the methylation level of the low methylation group in the previous and next iterative calculations are both less than 0.0001.

9. The method according to claim 1, characterized in that: According to the following criteria, the primary imprinted regulatory region and the adjacent primary window are extended to obtain the genomic imprinted-like regulatory region: (i) the distance between the primary imprinting-like regulatory region and the adjacent primary window is no more than 1000 bp, preferably no more than 800 bp, more preferably no more than 700 bp, and most preferably no more than 600 bp; (ii) the primary imprinting-like regulatory region and each of the adjacent primary windows contain at least 5 CpG sites, preferably at least 10 CpG sites; (iii) after extending the primary imprinted regulatory region and the adjacent primary window, the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.8, for example, not less than 0.85 or not less than 0.90; and (iv) After extending the primary imprinting regulatory region and the adjacent primary window, the number of sequencing reads in the high methylation group and the low methylation group with fewer sequencing reads accounts for no less than 30% of the total number of sequencing reads falling into the primary window, for example, no less than 40%.

10. The method according to claim 1, characterized in that Further including: Excluding the DNA methylation pattern region with tissue specificity from the genomic imprinting-like regulatory region to obtain the screened genomic imprinting-like regulatory region; and The screened genomic imprinted-like regulatory region is judged for similarity with the known imprinted regulatory region, so as to output the final genomic imprinted-like regulatory region.

11. The method according to claim 10, characterized in that The similarity determination is based on the following formula: p=1 / (1+e -(-31.33+20.2256*diff_score+0.0043*length+13.2152*ratio+0.0903*cg) ); in, p represents the similarity between the screened genomic imprinted regulatory region and the known imprinted regulatory region; diff_score represents the absolute value (diff_score) of the difference in methylation levels between the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region; cg represents the number of CpG sites in the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region; ratio represents the ratio of the number of sequencing reads of the high methylation group to the low methylation group in the screened genomic imprinting-like regulatory region; and Length represents the length of the candidate region in the screened genomic imprinting-like regulatory region.

12. A device for determining a genomic imprinting-like regulatory region, characterized in that: include: An acquisition module, used to acquire methylation sequencing data of an individual, wherein the methylation sequencing data is composed of a plurality of sequencing read segments; A primary window division module, used for dividing the genome reference sequence of the species corresponding to the individual into a plurality of primary windows, wherein the primary windows contain CpG sites; a primary imprinted-like regulatory region determination module, configured to select a primary imprinted-like regulatory region from the plurality of primary windows based on the methylation status of the CpG site in the primary window; and an optional genomic imprinting-like regulatory region determination module, configured to extend the primary imprinting-like regulatory region and the adjacent primary window to obtain the genomic imprinting-like regulatory region, in, The primary imprinted regulatory region is the primary window that meets the following criteria: (a) sequencing reads falling into the primary window can be classified into a high methylation group and a low methylation group, the methylation level of the high methylation group being higher than the methylation level of the low methylation group; and (b) The difference in the number of sequencing reads between the high methylation group and the low methylation group does not exceed a predetermined threshold.

13. The device according to claim 12, characterized in that The genomic imprinting-like regulatory region determination module is used to extend the primary imprinting regulatory region and the adjacent primary window according to the following criteria, so as to obtain the genomic imprinting-like regulatory region: (i) the distance between the primary imprinting-like regulatory region and the adjacent primary window is no more than 1000 bp, preferably no more than 800 bp, more preferably no more than 700 bp, and most preferably no more than 600 bp; (ii) the primary imprinting-like regulatory region and each of the adjacent primary windows contain at least 5 CpG sites, preferably at least 10 CpG sites; (iii) after extending the primary imprinted regulatory region and the adjacent primary window, the difference between the methylation level of the high methylation group and the methylation level of the low methylation group is not less than 0.8, for example, not less than 0.85 or not less than 0.90; and (iv) After extending the primary imprinting regulatory region and the adjacent primary window, the number of sequencing reads in the high methylation group and the low methylation group with fewer sequencing reads accounts for no less than 30% of the total number of sequencing reads falling into the primary window, for example, no less than 40%.

14. The device according to claim 12, characterized in that Further including: A screening module, used to exclude the DNA methylation pattern region with tissue specificity from the genomic imprinting-like regulatory region to obtain the screened genomic imprinting-like regulatory region; and The similarity judgment module is used to judge the similarity between the screened genomic imprinted-like regulatory region and the known imprinted regulatory region, so as to output the final genomic imprinted-like regulatory region.

15. The device according to claim 14, characterized in that The similarity determination is based on the following formula: p=1 / (1+e -(-31.33+20.2256*diff_score+0.0043*length+13.2152*ratio+0.0903*cg) ), in, p represents the similarity between the screened genomic imprinted regulatory region and the known imprinted regulatory region; diff_score represents the absolute value (diff_score) of the difference in methylation levels between the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region; cg represents the number of CpG sites in the high methylation group and the low methylation group in the screened genomic imprinting-like regulatory region; ratio represents the ratio of the number of sequencing reads of the high methylation group to the low methylation group in the screened genomic imprinting-like regulatory region; and Length represents the length of the candidate region in the screened genomic imprinting-like regulatory region.

16. A server, characterized in that: The server comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the method for determining a genomic imprinting-like regulatory region according to any one of claims 1 to 11 is implemented.

17. A computer-readable storage medium containing a computer program, characterized in that: When the computer program is executed by one or more processors, the method for determining a genomic imprinting-like regulatory region according to any one of claims 1 to 12 is implemented.

Images