{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,25]],"date-time":"2026-09-25T10:05:52Z","timestamp":1790330752035,"version":"4.1.0"},"reference-count":33,"publisher":"Oxford University Press (OUP)","issue":"9","license":[{"start":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T00:00:00Z","timestamp":1789516800000},"content-version":"vor","delay-in-days":16,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,8,31]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Motivation<\/jats:title>\n                    <jats:p>Accurate prediction of B-cell epitopes is essential for antibody design and vaccine development, yet remains fundamentally challenging. A major but often overlooked limitation of existing predictors is their implicit assumption that epitope identity can be decomposed into independent residue-level signals. In contrast, antibody recognition is governed by spatially contiguous surface patches whose functional identity emerges from coherent three-dimensional geometry rather than isolated residues.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Here, we reformulate structure-based B-cell epitope prediction as a surface-patch learning problem and introduce PatchEpi, a patch-aware geometric deep learning framework for antigen-intrinsic epitope modeling. It employs a patch-aware attention encoder and boundary-contrastive objectives to align the learning patch signal with the biological reality of antibody binding. PatchEpi integrates patch embeddings, fine-tuned ESM2 representations, and an equivariant message-passing network to model the 3D topology of antigen surfaces. To enable rigorous evaluation, we created a homology leakage-controlled split by reclustering the ANABAG dataset. Across multiple benchmarks, PatchEpi consistently outperforms residue-centric and graph-based state-of-the-art methods. Structural analyses further show that the model learns coherent surface patches that closely match experimentally resolved antibody footprints. These results demonstrate that accurate epitope prediction benefits primarily from reformulating the task around surface-patch learning. Patch-aware learning provides a principled and biologically grounded pathway toward more reliable structure-based epitope detection.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Availability and implementation<\/jats:title>\n                    <jats:p>The code and related resources are available at GitHub (https:\/\/github.com\/wsicheng739\/PatchEpi) and Zenodo (https:\/\/doi.org\/10.5281\/zenodo.20366444).<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btag678","type":"journal-article","created":{"date-parts":[[2026,9,14]],"date-time":"2026-09-14T11:13:07Z","timestamp":1789384387000},"source":"Crossref","is-referenced-by-count":0,"title":["PatchEpi: patch-aware equivariant learning improves structure-based epitope prediction"],"prefix":"10.1093","volume":"42","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-9992-9140","authenticated-orcid":false,"given":"Sicheng","family":"Wen","sequence":"first","affiliation":[{"name":"MARS Department, Viva Biotech (Shanghai) Limited , Shanghai 201318,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3685-7164","authenticated-orcid":false,"given":"Fei","family":"Li","sequence":"additional","affiliation":[{"name":"MARS Department, Viva Biotech (Shanghai) Limited , Shanghai 201318,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1989-7586","authenticated-orcid":false,"given":"Yue","family":"Qian","sequence":"additional","affiliation":[{"name":"MARS Department, Viva Biotech (Shanghai) Limited , Shanghai 201318,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2026,9,15]]},"reference":[{"key":"2026092505371034700_btag678-B1","doi-asserted-by":"publisher","first-page":"3031","DOI":"10.1021\/acs.jctc.7b00125","article-title":"The Rosetta All-Atom energy function for macromolecular modeling and design","volume":"13","author":"Alford","year":"2017","journal-title":"J Chem Theory Comput"},{"key":"2026092505371034700_btag678-B2","doi-asserted-by":"publisher","first-page":"2908","DOI":"10.1038\/s41598-020-59834-2","article-title":"Evaluating the significance of contact maps in low-homology protein modeling using contact-assisted threading","volume":"10","author":"Bhattacharya","year":"2020","journal-title":"Sci Rep"},{"key":"2026092505371034700_btag678-B3","doi-asserted-by":"publisher","first-page":"27496","DOI":"10.1038\/s41598-024-78506-z","article-title":"B cell epitope prediction by capturing spatial clustering property of the epitopes using graph attention network","volume":"14","author":"Choi","year":"2024","journal-title":"Sci Rep"},{"key":"2026092505371034700_btag678-B4","doi-asserted-by":"publisher","first-page":"bbac567","DOI":"10.1093\/bib\/bbac567","article-title":"Critical review of conformational B-cell epitope prediction methods","volume":"24","author":"Cia","year":"2023","journal-title":"Brief Bioinform"},{"key":"2026092505371034700_btag678-B5","doi-asserted-by":"publisher","first-page":"e4497","DOI":"10.1002\/pro.4497","article-title":"BepiPred-3.0: improved B-Cell epitope prediction using protein language models","volume":"31","author":"Clifford","year":"2022","journal-title":"Protein Sci"},{"key":"2026092505371034700_btag678-B6","doi-asserted-by":"publisher","first-page":"130674","DOI":"10.1016\/j.bbagen.2024.130674","article-title":"B cell epitope mapping: the journey to better vaccines and therapeutic antibodies","volume":"1868","author":"De Leon","year":"2024","journal-title":"Biochim Biophys Acta Gen Subj"},{"key":"2026092505371034700_btag678-B7","first-page":"9323"},{"key":"2026092505371034700_btag678-B8","doi-asserted-by":"publisher","first-page":"11454","DOI":"10.1021\/acs.jcim.5c01599","article-title":"ANABAG: annotated antibody\u2013antigen data set with unique features for antibody engineering applications","volume":"65","author":"Grandguillaume","year":"2025","journal-title":"J Chem Inf Model"},{"key":"2026092505371034700_btag678-B9","doi-asserted-by":"publisher","first-page":"1322712","DOI":"10.3389\/fimmu.2024.1322712","article-title":"DiscoTope-3.0: improved B-cell epitope prediction using inverse folding latent representations","volume":"15","author":"H\u00f8ie","year":"2024","journal-title":"Front Immunol"},{"key":"2026092505371034700_btag678-B10","article-title":"LoRA: low-rank adaptation of large language models","volume-title":"International Conference on Learning Representations (ICLR).","author":"Hu","year":"2022"},{"key":"2026092505371034700_btag678-B11","doi-asserted-by":"publisher","first-page":"W533","DOI":"10.1093\/nar\/gkae386","article-title":"SEMA 2.0: web-platform for B-cell conformational epitopes prediction using artificial intelligence","volume":"52","author":"Ivanisenko","year":"2024","journal-title":"Nucleic Acids Res"},{"key":"2026092505371034700_btag678-B12","doi-asserted-by":"publisher","first-page":"188802","DOI":"10.1016\/j.bbcan.2022.188802","article-title":"Interleukin 13 receptor alpha 2 (IL13R\u03b12): expression, signaling pathways and therapeutic applications in cancer","volume":"1877","author":"Ja\u00e9n","year":"2022","journal-title":"Biochim Biophys Acta Rev Cancer"},{"key":"2026092505371034700_btag678-B13","doi-asserted-by":"publisher","first-page":"888","DOI":"10.3389\/fimmu.2018.00888","article-title":"Tuning the cytokine responses: an update on interleukin (IL)-4 and IL-13 receptor complexes","volume":"9","author":"Junttila","year":"2018","journal-title":"Front Immunol"},{"key":"2026092505371034700_btag678-B14","first-page":"102.285","article-title":"Boundary loss for highly unbalanced segmentation","year":"2019"},{"key":"2026092505371034700_btag678-B15","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1016\/j.molimm.2012.06.001","article-title":"Structural analysis of B-cell epitopes in antibody: protein complexes","volume":"53","author":"Kringelum","year":"2013","journal-title":"Mol Immunol"},{"key":"2026092505371034700_btag678-B16","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2002.07031","article-title":"Regularizing semi-supervised graph convolutional networks with a manifold smoothness loss","author":"Li","year":"2020"},{"key":"2026092505371034700_btag678-B17","doi-asserted-by":"publisher","first-page":"1123","DOI":"10.1126\/science.ade2574","article-title":"Evolutionary-scale prediction of atomic-level protein structure with a language model","volume":"379","author":"Lin","year":"2023","journal-title":"Science"},{"key":"2026092505371034700_btag678-B18","doi-asserted-by":"publisher","first-page":"890943","DOI":"10.3389\/fimmu.2022.890943","article-title":"A structure-based B-cell epitope prediction model through combing local and global features","volume":"13","author":"Lu","year":"2022","journal-title":"Front Immunol"},{"key":"2026092505371034700_btag678-B19","doi-asserted-by":"publisher","first-page":"1034","DOI":"10.1016\/j.drudis.2020.03.006","article-title":"Epitope-based vaccine design: a comprehensive overview of bioinformatics approaches","volume":"25","author":"Parvizpour","year":"2020","journal-title":"Drug Discov Today"},{"key":"2026092505371034700_btag678-B20","doi-asserted-by":"publisher","first-page":"E2656","DOI":"10.1073\/pnas.1401131111","article-title":"Origins of specificity and affinity in antibody\u2013protein interactions","volume":"111","author":"Peng","year":"2014","journal-title":"Proc Natl Acad Sci USA"},{"key":"2026092505371034700_btag678-B21","doi-asserted-by":"publisher","first-page":"208","DOI":"10.1016\/j.jmb.2016.12.005","article-title":"Structural characterisation reveals mechanism of IL-13-Neutralising monoclonal antibody tralokinumab as inhibition of binding to IL-13R\u03b11 and IL-13R\u03b12","volume":"429","author":"Popovic","year":"2017","journal-title":"J Mol Biol"},{"key":"2026092505371034700_btag678-B22","doi-asserted-by":"publisher","first-page":"btac759","DOI":"10.1093\/bioinformatics\/btac759","article-title":"DeepRank-GNN: a graph neural network framework to learn patterns in protein\u2013protein interfaces","volume":"39","author":"R\u00e9au","year":"2023","journal-title":"Bioinformatics"},{"key":"2026092505371034700_btag678-B23"},{"key":"2026092505371034700_btag678-B24","doi-asserted-by":"publisher","first-page":"1026","DOI":"10.1038\/nbt.3988","article-title":"MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets","volume":"35","author":"Steinegger","year":"2017","journal-title":"Nat Biotechnol"},{"key":"2026092505371034700_btag678-B25","doi-asserted-by":"publisher","first-page":"730","DOI":"10.1038\/s41592-022-01490-7","article-title":"ScanNet: an interpretable geometric deep learning model for structure-based protein binding site prediction","volume":"19","author":"Tubiana","year":"2022","journal-title":"Nat Methods"},{"key":"2026092505371034700_btag678-B26","doi-asserted-by":"publisher","first-page":"1330","DOI":"10.1016\/j.jmb.2013.01.024","article-title":"Structural basis of signaling blockade by anti-IL-13 antibody lebrikizumab","volume":"425","author":"Ultsch","year":"2013","journal-title":"J Mol Biol"},{"key":"2026092505371034700_btag678-B27","doi-asserted-by":"publisher","first-page":"2159","DOI":"10.3390\/app15042159","article-title":"GraphEPN: a deep learning framework for B-cell epitope prediction leveraging graph neural networks","volume":"15","author":"Wang","year":"2025","journal-title":"Appl Sci"},{"key":"2026092505371034700_btag678-B28","doi-asserted-by":"publisher","first-page":"947","DOI":"10.1093\/bioinformatics\/btab761","article-title":"Deep graph learning of inter-protein contacts","volume":"38","author":"Xie","year":"2022","journal-title":"Bioinformatics"},{"key":"2026092505371034700_btag678-B29","doi-asserted-by":"publisher","first-page":"btag006","DOI":"10.1093\/bioinformatics\/btag006","article-title":"RoBep: a region-oriented deep learning model for B-cell epitope prediction","volume":"42","author":"Xu","year":"2026","journal-title":"Bioinformatics"},{"key":"2026092505371034700_btag678-B30","doi-asserted-by":"publisher","first-page":"104489","DOI":"10.1016\/j.drudis.2025.104489","article-title":"Improving B-cell epitope prediction","volume":"30","author":"Yu","year":"2025","journal-title":"Drug Discov Today"},{"key":"2026092505371034700_btag678-B31","doi-asserted-by":"publisher","first-page":"52","DOI":"10.3390\/antib12030052","article-title":"Recent progress in antibody epitope prediction","volume":"12","author":"Zeng","year":"2023","journal-title":"Antibodies"},{"key":"2026092505371034700_btag678-B32","doi-asserted-by":"publisher","first-page":"btad187","DOI":"10.1093\/bioinformatics\/btad187","article-title":"Identifying B-cell epitopes using AlphaFold2 predicted structures and pretrained language model","volume":"39","author":"Zeng","year":"2023","journal-title":"Bioinformatics"},{"key":"2026092505371034700_btag678-B33","doi-asserted-by":"publisher","first-page":"W388","DOI":"10.1093\/nar\/gkz413","article-title":"SEPPA 3.0\u2014enhanced spatial epitope prediction enabling glycoprotein antigens","volume":"47","author":"Zhou","year":"2019","journal-title":"Nucleic Acids Res"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btag678\/71110216\/btag678.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/9\/btag678\/71110216\/btag678.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/9\/btag678\/71110216\/btag678.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,25]],"date-time":"2026-09-25T09:37:18Z","timestamp":1790329038000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/doi\/10.1093\/bioinformatics\/btag678\/8800082"}},"subtitle":[],"editor":[{"given":"Lenore","family":"Cowen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,8,31]]},"references-count":33,"journal-issue":{"issue":"9","published-print":{"date-parts":[[2026,8,31]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btag678","relation":{},"ISSN":["1367-4803","1367-4811"],"issn-type":[{"value":"1367-4803","type":"print"},{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,9]]},"published":{"date-parts":[[2026,8,31]]},"article-number":"btag678"}}