{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,19]],"date-time":"2026-10-19T03:40:55Z","timestamp":1792381255933,"version":"4.4.0"},"reference-count":22,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,7,23]],"date-time":"2026-07-23T00:00:00Z","timestamp":1784764800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,10,19]],"date-time":"2026-10-19T00:00:00Z","timestamp":1792368000000},"content-version":"vor","delay-in-days":88,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Indonesian Endowment Fund for Education scholarships"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["BMC Bioinformatics"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Current methods for predicting gene expression from histone modifications rely on arbitrary binary classification thresholds such as the median to distinguish between high and low expression for genes. This approach lacks biological justification, creates dataset-dependent classifications, and ignores the relative regulatory relationships between genes that are often more biologically meaningful than absolute cutoffs.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We developed a deep learning method using pairwise ranking to determine the relative gene expression levels between gene pairs based on their histone modification signals. The approach adopted the DirectRanker architecture, which enforces antisymmetry by construction through a twin-subnet design with asymmetric subtraction, and was evaluated against three baseline classifiers (Random Forest, Logistic Regression, and SVM Linear) on identical features and data splits. Ablation testing across all 31 non-empty subsets of five histone marks (H3K4me3, H3K9ac, H3K9me3, H3K27ac, H3K27me3) was conducted to identify their individual contributions. Datasets were strictly partitioned (80% training, 10% validation, 10% testing) with gene-level separation to prevent data leakage, and all experiments were repeated across five independent random seeds. Four datasets were analysed: two normal adult liver cell datasets (Donor\u00a03 and Donor\u00a04 from NCBI GEO GSE19465) and two HepG2 hepatocellular carcinoma cell line datasets (ENCSR134DWG and GSE76344).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Multi-mark combinations anchored by active histone marks (H3K27ac, H3K4me3, H3K9ac) consistently drove predictive performance, with the best combinations achieving AUROC values of 0.833\u22120.867 and test accuracies of 74\u201378% across datasets. DirectRanker attained AUPRC values of 0.819\u22120.856, closely matching or exceeding Random Forest on precision\u2013recall performance. Repressive marks (H3K9me3, H3K27me3) consistently underperformed when used in isolation, with single-mark models achieving test accuracies of only 50\u201366%. DirectRanker produced substantially higher antisymmetry scores (0.961\u22120.983) than all baseline classifiers (0.787\u22120.851), confirming logically consistent pairwise predictions. Performance plateaued at four- to five-mark combinations, reflecting correlated predictive signal among the active marks within the promoter-proximal feature space.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>The pairwise ranking framework provides a principled alternative to threshold-based classification, enabling relative gene expression comparisons without arbitrary binarisation cutoffs. While Random Forest achieved marginally higher accuracy, DirectRanker\u2019s structural antisymmetry guarantee makes it the more suitable choice for applications requiring a globally coherent gene ranking. Active promoter marks carry the majority of the predictive signal, while the performance saturation with increasing mark combinations reflects overlapping information content rather than biological redundancy.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1186\/s12859-026-06553-6","type":"journal-article","created":{"date-parts":[[2026,7,23]],"date-time":"2026-07-23T07:44:22Z","timestamp":1784792662000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["A deep learning framework for threshold-free relative gene expression ranking from histone modification signals"],"prefix":"10.1186","volume":"27","author":[{"given":"Abdul","family":"Munif","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Amitava","family":"Datta","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhaoyu","family":"Li","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Max","family":"Ward","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,7,23]]},"reference":[{"key":"6553_CR1","doi-asserted-by":"publisher","first-page":"160","DOI":"10.1289\/EHP.114-A160","volume":"114","author":"B Weinhold","year":"2006","unstructured":"Weinhold B. Epigenetics: The science of change. Environ Health Perspect. 2006;114:160. https:\/\/doi.org\/10.1289\/EHP.114-A160.","journal-title":"Environ Health Perspect"},{"key":"6553_CR2","doi-asserted-by":"publisher","first-page":"676","DOI":"10.1016\/J.TIG.2022.03.010","volume":"38","author":"AL Mattei","year":"2022","unstructured":"Mattei AL, Bailly N, Meissner A. Dna methylation: a historical perspective. Trends Genet. 2022;38:676\u2013707. https:\/\/doi.org\/10.1016\/J.TIG.2022.03.010.","journal-title":"Trends Genet"},{"key":"6553_CR3","doi-asserted-by":"publisher","first-page":"865","DOI":"10.1007\/S10739-022-09691-8\/FIGURES\/1","volume":"55","author":"JD Tompkins","year":"2022","unstructured":"Tompkins JD. Discovering dna methylation, the history and future of the writing on dna. J Hist Biol. 2022;55:865\u201387. https:\/\/doi.org\/10.1007\/S10739-022-09691-8\/FIGURES\/1.","journal-title":"J Hist Biol"},{"key":"6553_CR4","doi-asserted-by":"publisher","first-page":"639","DOI":"10.1093\/BIOINFORMATICS\/BTW427","volume":"32","author":"R Singh","year":"2016","unstructured":"Singh R, Lanchantin J, Robins G, Qi Y. Deepchrome: deep-learning for predicting gene expression from histone modifications. Bioinformatics. 2016;32:639\u201348. https:\/\/doi.org\/10.1093\/BIOINFORMATICS\/BTW427.","journal-title":"Bioinformatics"},{"key":"6553_CR5","doi-asserted-by":"publisher","unstructured":"Kadavath S, Paradis S, Yeung J. Deepchrome 2.0: Investigating and improving architectures, visualizations, & experiments (2022). https:\/\/doi.org\/10.48550\/arXiv.2209.11923","DOI":"10.48550\/arXiv.2209.11923"},{"key":"6553_CR6","doi-asserted-by":"publisher","unstructured":"Burges C, Shaked T, Renshaw E, Lazier A, Deeds M, Hamilton N, Hullender G. Learning to rank using gradient descent. In: Proceedings of the 22nd International Conference on Machine Learning. ICML \u201905, pp. 89\u201396. Association for Computing Machinery, New York, NY, USA (2005). https:\/\/doi.org\/10.1145\/1102351.1102363. https:\/\/doi.org\/10.1145\/1102351.1102363","DOI":"10.1145\/1102351.1102363"},{"key":"6553_CR7","unstructured":"Burges CJC. From ranknet to lambdarank to lambdamart: An overview. Technical Report MSR-TR-2010-82 (June 2010). https:\/\/www.microsoft.com\/en-us\/research\/publication\/from-ranknet-to-lambdarank-to-lambdamart-an-overview\/"},{"key":"6553_CR8","doi-asserted-by":"publisher","first-page":"237","DOI":"10.1007\/978-3-030-46133-1_15","volume-title":"Machine Learning and Knowledge Discovery in Databases","author":"M K\u00f6ppel","year":"2020","unstructured":"K\u00f6ppel M, Segner A, Wagener M, Pensel L, Karwath A, Kramer S. Pairwise learning to rank by neural networks revisited: Reconstruction, theoretical analysis and practical performance. In: Brefeld U, Fromont E, Hotho A, Knobbe A, Maathuis M, Robardet C, editors. Machine Learning and Knowledge Discovery in Databases. Cham: Springer; 2020. p. 237\u201352."},{"key":"6553_CR9","doi-asserted-by":"publisher","first-page":"151","DOI":"10.1186\/S12859-022-04687-X","volume":"23","author":"F Frasca","year":"2022","unstructured":"Frasca F, Matteucci M, Leone M, Morelli MJ, Masseroli M. Accurate and highly interpretable prediction of gene expression from histone modifications. BMC Bioinform. 2022;23:151. https:\/\/doi.org\/10.1186\/S12859-022-04687-X.","journal-title":"BMC Bioinform"},{"key":"6553_CR10","first-page":"6785","volume":"30","author":"R Singh","year":"2017","unstructured":"Singh R, Lanchantin J, Sekhon A, Qi Y. Attend and predict: Understanding gene regulation by selective attention on chromatin. Adv Neural Inf Process Syst. 2017;30:6785.","journal-title":"Adv Neural Inf Process Syst"},{"key":"6553_CR11","doi-asserted-by":"publisher","first-page":"1081842","DOI":"10.3389\/FGENE.2022.1081842\/BIBTEX","volume":"13","author":"Y Chen","year":"2022","unstructured":"Chen Y, Xie M, Wen J. Predicting gene expression from histone modifications with self-attention based neural networks and transfer learning. Front Genet. 2022;13:1081842. https:\/\/doi.org\/10.3389\/FGENE.2022.1081842\/BIBTEX.","journal-title":"Front Genet"},{"key":"6553_CR12","doi-asserted-by":"publisher","DOI":"10.1016\/J.NEUCOM.2024.128645","volume":"610","author":"H Hu","year":"2024","unstructured":"Hu H, Wang X, Zhang Y, Chen Q, Guan Q. A comprehensive survey on contrastive learning. Neurocomputing. 2024;610:128645. https:\/\/doi.org\/10.1016\/J.NEUCOM.2024.128645.","journal-title":"Neurocomputing"},{"key":"6553_CR13","unstructured":"Chen T, Kornblith S, Norouzi M, Hinton G. A simple framework for contrastive learning of visual representations. 37th International Conference on Machine Learning, ICML 2020 PartF168147-3: 1575\u20131585 (2020)."},{"key":"6553_CR14","unstructured":"Khosla P, Teterwak P, Wang C, Sarna A, Research G, Tian Y, et al. 2020. Supervised contrastive learning. Adv neural inform process syst. 33: 18661-18673."},{"key":"6553_CR15","unstructured":"ENCODE Project Consortium: ENCDO776UMN: Human donor (25 years, female, European) metadata. https:\/\/www.encodeproject.org\/human-donors\/ENCDO776UMN\/. Accessed: 2026-02-05 (2026)."},{"key":"6553_CR16","unstructured":"ENCODE Project Consortium: ENCDO832GKP: Human donor (31 years, male, healthy) metadata. https:\/\/www.encodeproject.org\/human-donors\/ENCDO832GKP\/. Accessed: 2026-02-05 (2026)."},{"key":"6553_CR17","unstructured":"National Center for Biotechnology Information (NCBI) Gene Expression Omnibus: GSE76344: Histone modification and nucleosome mapping in human liver cancer cells. https:\/\/www.ncbi.nlm.nih.gov\/geo\/query\/acc.cgi?acc=GSE76344. Accessed: 2026-02-05 (2020)."},{"key":"6553_CR18","unstructured":"ENCODE Project Consortium: ENCSR134DWG: Reference epigenome collection (HepG2 histone\/chromatin data) metadata. https:\/\/www.encodeproject.org\/reference-epigenomes\/ENCSR134DWG\/. Accessed: 2026-02-05 (2026)."},{"key":"6553_CR19","unstructured":"Primary data processing and quality control \u2014 processed data from the NIH Roadmap Epigenomics Mapping Consortium. https:\/\/egg2.wustl.edu\/roadmap\/web_portal\/processed_data.html. Accessed: 2025-07-01 (2015)."},{"key":"6553_CR20","unstructured":"Kundaje Lab: PhantomPeakQualTools: Quality Metrics for ChIP-seq and Related Sequencing Assays. Bioconda \/ Anaconda.org. Conda package on Bioconda; accessed 2026-02-05. https:\/\/anaconda.org\/bioconda\/phantompeakqualtools"},{"key":"6553_CR21","unstructured":"Munif A. munif\/epi-thesis. https:\/\/github.com\/munif\/epi-thesis\/. Accessed: 2025-11-15 (2025)."},{"key":"6553_CR22","unstructured":"Infrastructure and Platforms: KAYA research compute platform. The University of Western Australia. Accessed: 2026\u201302-05 (2025). https:\/\/doi.org\/10.26182\/DH5S-V482"}],"container-title":["BMC Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12859-026-06553-6","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12859-026-06553-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12859-026-06553-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,19]],"date-time":"2026-10-19T03:16:06Z","timestamp":1792379766000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12859-026-06553-6"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,23]]},"references-count":22,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["6553"],"URL":"https:\/\/doi.org\/10.1186\/s12859-026-06553-6","relation":{},"ISSN":["1471-2105"],"issn-type":[{"value":"1471-2105","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,23]]},"assertion":[{"value":"2 February 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"3 July 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"23 July 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Not applicable.","order":1,"name":"Ethics","label":"Ethics approval and consent to participate","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable.","order":2,"name":"Ethics","label":"Consent for publication","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":3,"name":"Ethics","label":"Competing Interests","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"240"}}