[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-367069-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-367069-en":130},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","methylbert-enables-read-level-dna-methylation-pattern-identification-and-tumour-deconvolution-using-a-transformer-based-model","MethylBERT enables read-level DNA methylation pattern identification and tumour deconvolution using a Transformer-based model","","DNA methylation (DNAm) is a major epigenetic mark whose cancer-related changes motivate better analysis beyond summarized array-based signals. MethylBERT proposes a Transformer-based model that operates at read level by learning methylation patterns together with local genomic sequence to classify reads as tumour-derived or normal and estimate tumour cell fractions from bulk samples. Evaluation shows improved deconvolution performance across read properties, and demonstrated usefulness for cell-type deconvolution and non-invasive early cancer diagnostics with liquid biopsy data.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/methylbert-enables-read-level-dna-methylation-pattern-identification-and-tumour-deconvolution-using-a-transformer-based-model/367069/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/methylbert-enables-read-level-dna-methylation-pattern-identification-and-tumour-deconvolution-using-a-transformer-based-model/367069.png","ImageObject",300,407,{"name":42,"@type":43},"Patrick","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-11-06","2026-09-23",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",19,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What does MethylBERT do at the read level?","Question",{"text":63,"@type":64},"MethylBERT classifies sequence reads into tumour or normal cell types using learned methylation patterns and local genomic sequence, then derives tumour purity estimates from the resulting posterior probabilities.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"How does MethylBERT estimate tumour cell fractions in bulk samples?",{"text":68,"@type":64},"It converts posterior cell-type probabilities into tumour purity estimates using Bayesian probability inversion and maximum likelihood estimation, with estimation adjustment informed by region-wise tumour purity.",{"name":70,"@type":61,"acceptedAnswer":71},"What evidence shows MethylBERT’s advantages and where can it be applied?",{"text":72,"@type":64},"The evaluation reports better performance than existing deconvolution methods across varying methylation pattern complexity, read length, and coverage, and it also supports cell-type deconvolution and non-invasive early cancer diagnostics from liquid biopsy (ctDNA) samples.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},367069,1790616935,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,115,120,123,127],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":112,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":113,"slug":114},8,30,"research-report",{"id":116,"doc_module":4,"doc_module_name":25,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":25,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":25,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":55,"doc_module":4,"doc_module_name":25,"category_name":128,"show_sort_weight":97,"slug":129},"General","general",{"code":4,"msg":82,"data":131},{"doc_id":79,"user_id":132,"nickname":42,"user_avatar":133,"doc_module":4,"category_id":112,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":139,"language":140,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":12,"update_tm":144,"read_time":145,"preview_page_count":33,"preview_urls":146},549758146520,"https://ap-avatar.wpscdn.com/avatar/80002397d8c0411e94?_k=1775819394049821470","Article [https://doi.org/10.1038/s41467-025-55920-z](https://doi.org/10.1038/s41467-025-55920-z)  \nMethylBERT enables read-level DNA methylation pattern identiﬁcation and tumour deconvolution using a Transformerbased model  \nReceived: 4 May 2024  \n\n| Accepted: 2 January 2025 |\n| --- |\n| |\n| Check for updates |\n\nYunhee Jeong 1 , Clarissa Gerhäuser 1, Guido Sauter2, Thorsten Schlomm 3, Karl Rohr4 & Pavlo Lutsik 1,5   \nDNA methylation (DNAm) is a key epigenetic mark that shows profound alterations in cancer. Read-level methylomes enable more in-depth analyses, due to their broad genomic coverage and preservation of rare cell-type signals, compared to summarized data such as 450K/EPIC microarrays. Here, we propose MethylBERT, a Transformer-based model for read-level methylation pattern classiﬁcation. MethylBERT identiﬁes tumour-derived sequence reads based on their methylation patterns and local genomic sequence, and estimates tumour cell fractions within bulk samples. In our evaluation, MethylBERT outperforms existing deconvolution methods and demonstrates high accuracy regardless of methylation pattern complexity, read length and read coverage. Moreover, we show its applicability to cell-type deconvolution as well as non-invasive early cancer diagnostics using liquid biopsy samples. MethylBERT represents a signiﬁcant advancement in read-level methylome analysis and enables accurate tumour purity estimation. The broad applicability of MethylBERT will enhance studies on both tumour and non-cancerous bulk methylomes.  \nDNA methylation (DNAm) refers to enzymatic modiﬁcation of nucleotide bases in the DNA with methyl groups. In animal and many plant genomes, cytosine followed by guanine (CpG) is the most frequent methylation target. Aberrant methylation patterns at CpGs reﬂect epigenetic heterogeneity in human tumours1. Therefore, DNAm data has been widely used for estimating tumour purity to examine the tumour epigenetic landscape relevant to clinical outcomes, tumour diagnosis and phenotypic characteristics2–4.  \nDNAm can be proﬁled by sequencing methods such as whole genome bisulﬁte sequencing (WGBS)5 or reduced representation bisulﬁte sequencing6, ampliﬁcation-free long-read sequencing, e.g. using  \nOxford Nanopore Technologies or Paciﬁc Biosciences platforms, as well as with bisulﬁte-based microarray methods like Inﬁnium 450K/EPIC arrays7. High-quality sequencing-based proﬁling produces sequence reads covering a broad range of genomic regions with sufﬁcient read depth (30x being a de facto industry-wide standard), and thereby preserves single-molecule signals of rare cell populations8. Sequencingbased data is even more crucial in circulating tumour DNA (ctDNA) analysis that facilitates non-invasive early diagnosis, prognosis and treatment response monitoring in cancer patients9–11. Nevertheless, most purity estimation or cell-type deconvolution methods have been developed for array-based DNAm proﬁles because of the intuitive  \n1Division of Cancer Epigenomics, German Cancer Research Center (DKFZ), Heidelberg, Germany. 2Institute for Pathology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. 3Department of Urology, Charité – Universitätsmedizin Berlin, Berlin, Germany. 4Biomedical Computer Vision Group, BioQuant, IPMB, Heidelberg University, Heidelberg, Germany. 5Department of Oncology, KU Leuven, Leuven, Belgium.  \ne-mail: y.jeong@dkfz-heidelberg.de; [pavlo.lutsik@kuleuven.be](pavlo.lutsik@kuleuven.be)  \napplication of classical linear algebraic algorithms to matrices of average DNAm levels (beta-values)12. Most ctDNA data analysis methods likewise use beta-values. Moreover, our previous analysis showed that existing sequencing-based deconvolution methods do not perform better than array-based methods implying that they fail to fully exploit the advantages of sequencing data for accurate inference13.  \nTo overcome these limitations, we propose MethylBERT, a deep learning method for read-level methylation pattern id","cbCaieUbJsMOzQe3","https://ap.wps.com/l/cbCaieUbJsMOzQe3","pdf",3523817,14,"English","# Introduction\n## DNA methylation and cancer relevance\n## Limitations of existing deconvolution approaches\n# Method\n## MethylBERT overview\n## Read-level methylation pattern identification and purity estimation\n# Results\n## Performance evaluation and model interpretability\n## Fisher information for precision guidance\n# Applications\n## Early cancer diagnostics using ctDNA\n## Cell-type deconvolution","[{\"question\":\"What does MethylBERT do at the read level?\",\"answer\":\"MethylBERT classifies sequence reads into tumour or normal cell types using learned methylation patterns and local genomic sequence, then derives tumour purity estimates from the resulting posterior probabilities.\"},{\"question\":\"How does MethylBERT estimate tumour cell fractions in bulk samples?\",\"answer\":\"It converts posterior cell-type probabilities into tumour purity estimates using Bayesian probability inversion and maximum likelihood estimation, with estimation adjustment informed by region-wise tumour purity.\"},{\"question\":\"What evidence shows MethylBERT’s advantages and where can it be applied?\",\"answer\":\"The evaluation reports better performance than existing deconvolution methods across varying methylation pattern complexity, read length, and coverage, and it also supports cell-type deconvolution and non-invasive early cancer diagnostics from liquid biopsy (ctDNA) samples.\"}]","MethylBERT enables read-level DNA methylation pattern identification and tumour deconvolution using a Transformer-based model | PDF",1790169130,35,[147,151,155,159],{"page_no":22,"url":148,"width":149,"height":150},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/970ddb19c378995057241c1b1f5d6f1a.webp",983,1305,{"page_no":26,"url":152,"width":153,"height":154},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/fedac2826c5b018dcf2d60692ae0c839.webp",982,1304,{"page_no":30,"url":156,"width":157,"height":158},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/3d196a719937f2ce8f8cab99f2612467.webp",981,1303,{"page_no":33,"url":160,"width":161,"height":162},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/b2abfb572fec50367cd6a36d4815daa7.webp",980,1301]