{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T04:51:16Z","timestamp":1791262276463,"version":"4.1.0"},"reference-count":0,"publisher":"Oxford University Press (OUP)","license":[{"start":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T00:00:00Z","timestamp":1791244800000},"content-version":"am","delay-in-days":1,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"Abstract<\/jats:title>\n \n Motivation<\/jats:title>\n Accurate identification of antibody paratopes\u2014the residues contacting the antigen\u2014is vital for understanding antibody\u2013antigen recognition, antibody engineering, and vaccine design. However, structure-based predictors require antibody structural inputs, which are often obtained from experimentally determined antibody\u2013antigen complexes in benchmark settings but are generally unavailable for new sequences. Conversely, sequence-based predictors scale efficiently but remain largely antigen-agnostic, limiting their use of pair-specific antigen context.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n We introduce ParaEM, an antigen-aware sequence-based paratope predictor that operates directly on antibody and antigen sequences. ParaEM leverages pretrained ESM3 embeddings to encode both sequences, augmenting antibody residues with trainable, CDR-informed embeddings to introduce a biologically informed inductive bias. It models interactions via a residue-level compatibility scorer that defines latent antigen-anchor distributions. During training, a generalized Expectation\u2013Maximization (EM) framework alternates between estimating label-conditioned posterior responsibilities (E-step) and performing gradient-based parameter updates (M-step). Evaluated across three established benchmarks, ParaEM outperforms existing sequence-based baselines in AUC-PR and achieves performance competitive with structure-based predictors without requiring explicit 3D structural inputs. Ablation studies show consistent improvements from generalized EM optimization and an additional gain from CDR-informed embeddings. Antibody-side and antigen-information controls further show that antibody sequence carries substantial predictive signal, while the matching antigen provides additional predictive information. Together, these results support ParaEM as a scalable framework for incorporating matching-antigen context into sequence-based paratope prediction.<\/jats:p>\n <\/jats:sec>\n \n Availability and Implementation<\/jats:title>\n ParaEM is available at https:\/\/github.com\/kimtaegyuu\/ParaEM. An archival snapshot of the software corresponding to this study is available on Zenodo at DOI: 10.5281\/zenodo.22157100.<\/jats:p>\n <\/jats:sec>\n \n Supplementary information<\/jats:title>\n Supplementary data are available.<\/jats:p>\n <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btag739","type":"journal-article","created":{"date-parts":[[2026,10,2]],"date-time":"2026-10-02T11:45:42Z","timestamp":1790941542000},"source":"Crossref","is-referenced-by-count":0,"title":["ParaEM: Antigen-aware sequence-based paratope prediction with Expectation-Maximization"],"prefix":"10.1093","author":[{"given":"Taegyu","family":"Kim","sequence":"first","affiliation":[{"name":"Seoul National University Interdisciplinary Program in Artificial Intelligence, , 1, Gwanak-ro, 08826, Seoul, Republic of","place":["Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Changyun","family":"Cho","sequence":"additional","affiliation":[{"name":"AIGENDRUG Co., Ltd ., Gwanak-ro 1, Gwanak-gu, Seoul, 08826, South","place":["Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Inuk","family":"Jung","sequence":"additional","affiliation":[{"name":"Kyungpook National University, 80 Daegu University-ro School of Computer Science and Engineering, , Daegu, Republic of","place":["Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sun","family":"Kim","sequence":"additional","affiliation":[{"name":"Seoul National University Interdisciplinary Program in Artificial Intelligence, , 1, Gwanak-ro, 08826, Seoul, Republic of","place":["Korea"]},{"name":"Seoul National University Interdisciplinary Program in Bioinformatics, , 1, Gwanak-ro, 08826, Seoul, Republic of","place":["Korea"]},{"name":"Seoul National University Department of Computer Science and Engineering, , 1, Gwanak-ro, 08826, Seoul, Republic of","place":["Korea"]},{"name":"AIGENDRUG Co., Ltd ., Gwanak-ro 1, Gwanak-gu, Seoul, 08826, South","place":["Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2026,10,5]]},"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btag739\/71610665\/btag739.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btag739\/71610665\/btag739.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T04:03:25Z","timestamp":1791259405000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article\/doi\/10.1093\/bioinformatics\/btag739\/8866641"}},"subtitle":[],"editor":[{"given":"Can","family":"Alkan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,10,5]]},"references-count":0,"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btag739","relation":{},"ISSN":["1367-4811"],"issn-type":[{"value":"1367-4811","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,5]]},"article-number":"btag739"}}