{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,24]],"date-time":"2026-04-24T15:15:38Z","timestamp":1777043738225,"version":"3.51.4"},"reference-count":29,"publisher":"Oxford University Press (OUP)","issue":"4","license":[{"start":{"date-parts":[[2026,3,14]],"date-time":"2026-03-14T00:00:00Z","timestamp":1773446400000},"content-version":"vor","delay-in-days":1,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"ANPCyT PICT 2022"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,4,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Motivation<\/jats:title>\n                    <jats:p>Due to the rapid growth of sequence generation, which has surpassed the expert curators ability to manually review and annotate them, the computational annotation of proteins remains a significant challenge in bioinformatics nowadays. The Pfam database contains a large collection of proteins that are annotated with domain families through profile Hidden Markov models (pHMMs). Using the aligned sequences of a curated family, one HMM is trained independently for each family, missing the opportunity of learning patterns across families, i.e. from a complete view of all the dataset. As an alternative, some deep learning (DL) models have been recently proposed, nevertheless with simple representations of the inputs and moderate improvements in performance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>In this work, we present ET-Pfam, a novel approach based on transfer learning and ensembles of multiple DL classifiers to predict functional families in the Pfam database. Several base DL models are first trained using learned representations from protein large language models. Then, the base models are integrated using classical ensemble strategies and novel voting approaches by learning weights for each model and for each Pfam family. Results demonstrate that the proposed ET-Pfam method can consistently diminish error rates compared to individual DL models, boosting prediction performance. Among the novel ensemble strategies presented here, the learned weights by family voting achieved the best performance, with the lowest error rate (7.00%), significantly surpassing the best individual base model error (12.91%) and competitors of the state-of-the-art.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Availability and implementation<\/jats:title>\n                    <jats:p>Data and source code are available at https:\/\/github.com\/sinc-lab\/ET-Pfam.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btag121","type":"journal-article","created":{"date-parts":[[2026,3,11]],"date-time":"2026-03-11T12:48:24Z","timestamp":1773233304000},"source":"Crossref","is-referenced-by-count":0,"title":["ET-Pfam: ensemble transfer learning for protein family prediction"],"prefix":"10.1093","volume":"42","author":[{"ORCID":"https:\/\/orcid.org\/0009-0007-5759-1058","authenticated-orcid":false,"given":"Sofia A","family":"Duarte","sequence":"first","affiliation":[{"name":"Research Institute for Signals, Systems and Computational Intelligence, sinc(i), FICH-UNL, CONICET, Ciudad Universitaria UNL , Santa Fe 3000,","place":["Argentina"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Rosario","family":"Vitale","sequence":"additional","affiliation":[{"name":"Research Institute for Signals, Systems and Computational Intelligence, sinc(i), FICH-UNL, CONICET, Ciudad Universitaria UNL , Santa Fe 3000,","place":["Argentina"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sofia","family":"Escudero","sequence":"additional","affiliation":[{"name":"Research Institute for Signals, Systems and Computational Intelligence, sinc(i), FICH-UNL, CONICET, Ciudad Universitaria UNL , Santa Fe 3000,","place":["Argentina"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Emilio","family":"Fenoy","sequence":"additional","affiliation":[{"name":"Research Institute for Signals, Systems and Computational Intelligence, sinc(i), FICH-UNL, CONICET, Ciudad Universitaria UNL , Santa Fe 3000,","place":["Argentina"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5702-946X","authenticated-orcid":false,"given":"Leandro A","family":"Bugnon","sequence":"additional","affiliation":[{"name":"Research Institute for Signals, Systems and Computational Intelligence, sinc(i), FICH-UNL, CONICET, Ciudad Universitaria UNL , Santa Fe 3000,","place":["Argentina"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2182-4351","authenticated-orcid":false,"given":"Diego H","family":"Milone","sequence":"additional","affiliation":[{"name":"Research Institute for Signals, Systems and Computational Intelligence, sinc(i), FICH-UNL, CONICET, Ciudad Universitaria UNL , Santa Fe 3000,","place":["Argentina"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4459-4560","authenticated-orcid":false,"given":"Georgina","family":"Stegmayer","sequence":"additional","affiliation":[{"name":"Research Institute for Signals, Systems and Computational Intelligence, sinc(i), FICH-UNL, CONICET, Ciudad Universitaria UNL , Santa Fe 3000,","place":["Argentina"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2026,3,13]]},"reference":[{"key":"2026042409465017700_btag121-B1","doi-asserted-by":"crossref","first-page":"3389","DOI":"10.1093\/nar\/25.17.3389","article-title":"Gapped BLAST and PSI-BLAST: a new generation of protein database search programs","volume":"25","author":"Altschul","year":"1997","journal-title":"Nucleic Acids Res"},{"key":"2026042409465017700_btag121-B2","doi-asserted-by":"crossref","first-page":"138D","DOI":"10.1093\/nar\/gkh121","article-title":"The pfam protein families database","volume":"32","author":"Bateman","year":"2004","journal-title":"Nucleic Acids Res"},{"key":"2026042409465017700_btag121-B3","doi-asserted-by":"crossref","first-page":"932","DOI":"10.1038\/s41587-021-01179-w","article-title":"Using deep learning to annotate the protein universe","volume":"40","author":"Bileschi","year":"2022","journal-title":"Nat Biotechnol"},{"key":"2026042409465017700_btag121-B4","doi-asserted-by":"crossref","first-page":"100691","DOI":"10.1016\/j.patter.2023.100691","article-title":"Transfer learning: the key to functionally annotate the protein universe","volume":"4","author":"Bugnon","year":"2023","journal-title":"Patterns (N Y)"},{"key":"2026042409465017700_btag121-B5","doi-asserted-by":"crossref","first-page":"500","DOI":"10.1038\/s42256-020-0217-y","article-title":"Ensemble deep learning in bioinformatics","volume":"2","author":"Cao","year":"2020","journal-title":"Nat Mach Intell"},{"key":"2026042409465017700_btag121-B6","doi-asserted-by":"crossref","first-page":"1914","DOI":"10.1038\/s41467-022-29443-w","article-title":"Learning meaningful representations of protein sequences","volume":"13","author":"Detlefsen","year":"2022","journal-title":"Nat Commun"},{"key":"2026042409465017700_btag121-B7","doi-asserted-by":"crossref","first-page":"755","DOI":"10.1093\/bioinformatics\/14.9.755","article-title":"Profile hidden markov models","volume":"14","author":"Eddy","year":"1998","journal-title":"Bioinformatics"},{"key":"2026042409465017700_btag121-B8","doi-asserted-by":"crossref","first-page":"e1002195","DOI":"10.1371\/journal.pcbi.1002195","article-title":"Accelerated profile hmm searches","volume":"7","author":"Eddy","year":"2011","journal-title":"PLoS Comput Biol"},{"key":"2026042409465017700_btag121-B9","doi-asserted-by":"crossref","first-page":"113","DOI":"10.1186\/1471-2105-5-113","article-title":"Muscle: a multiple sequence alignment method with reduced time and space complexity","volume":"5","author":"Edgar","year":"2004","journal-title":"BMC Bioinformatics"},{"key":"2026042409465017700_btag121-B10","doi-asserted-by":"crossref","first-page":"7112","DOI":"10.1109\/TPAMI.2021.3095381","article-title":"Prottrans: toward understanding the language of life through self-supervised learning","volume":"44","author":"Elnaggar","year":"2022","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"2026042409465017700_btag121-B11","doi-asserted-by":"crossref","first-page":"bbac232","DOI":"10.1093\/bib\/bbac232","article-title":"Transfer learning in proteins: evaluating novel protein learned representations for bioinformatics tasks","volume":"23","author":"Fenoy","year":"2022","journal-title":"Brief Bioinform"},{"key":"2026042409465017700_btag121-B12","doi-asserted-by":"crossref","first-page":"105151","DOI":"10.1016\/j.engappai.2022.105151","article-title":"Ensemble deep learning: a review","volume":"115","author":"Ganaie","year":"2022","journal-title":"Eng Appl Artif Intell"},{"key":"2026042409465017700_btag121-B13","doi-asserted-by":"crossref","first-page":"102986","DOI":"10.1016\/j.sbi.2025.102986","article-title":"Teaching ai to speak protein","volume":"91","author":"Heinzinger","year":"2025","journal-title":"Curr Opin Struct Biol"},{"key":"2026042409465017700_btag121-B14","first-page":"1","article-title":"LoRA: low-rank adaptation of large language models","volume":"1","author":"Hu","year":"2022","journal-title":"ICLR Proc"},{"key":"2026042409465017700_btag121-B15","doi-asserted-by":"crossref","first-page":"171","DOI":"10.1007\/s10930-024-10181-5","article-title":"A comprehensive review on machine learning techniques for protein family prediction","volume":"43","author":"Idhaya","year":"2024","journal-title":"Protein J"},{"key":"2026042409465017700_btag121-B16","doi-asserted-by":"crossref","first-page":"583","DOI":"10.1038\/s41586-021-03819-2","article-title":"Highly accurate protein structure prediction with alphafold","volume":"596","author":"Jumper","year":"2021","journal-title":"Nature"},{"key":"2026042409465017700_btag121-B17","doi-asserted-by":"crossref","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":"2026042409465017700_btag121-B18","doi-asserted-by":"crossref","first-page":"D523","DOI":"10.1093\/nar\/gkae997","article-title":"The pfam protein families database: embracing ai\/ml","volume":"53","author":"Paysan-Lafosse","year":"2025","journal-title":"Nucleic Acids Research"},{"key":"2026042409465017700_btag121-B19","doi-asserted-by":"crossref","first-page":"120","DOI":"10.1016\/j.ins.2016.04.027","article-title":"Diversity control for improving the analysis of consensus clustering","volume":"361\u2013362","author":"Pividori","year":"2016","journal-title":"Information Sciences"},{"key":"2026042409465017700_btag121-B20","doi-asserted-by":"crossref","first-page":"e2016239118","DOI":"10.1073\/pnas.2016239118","article-title":"Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences","volume":"118","author":"Rives","year":"2021","journal-title":"Proc Natl Acad Sci USA"},{"key":"2026042409465017700_btag121-B21","doi-asserted-by":"crossref","first-page":"7407","DOI":"10.1038\/s41467-024-51844-2","article-title":"Fine-tuning protein language models boosts predictions across diverse tasks","volume":"15","author":"Schmirler","year":"2024","journal-title":"Nat Commun"},{"key":"2026042409465017700_btag121-B22","doi-asserted-by":"crossref","first-page":"i254","DOI":"10.1093\/bioinformatics\/bty275","article-title":"Deepfam: deep learning based alignment-free method for protein family modeling and prediction","volume":"34","author":"Seo","year":"2018","journal-title":"Bioinformatics"},{"key":"2026042409465017700_btag121-B23","doi-asserted-by":"crossref","first-page":"3775","DOI":"10.3390\/ijms24043775","article-title":"Survey of protein sequence embedding models","volume":"24","author":"Tran","year":"2023","journal-title":"Int J Mol Sci"},{"key":"2026042409465017700_btag121-B24","doi-asserted-by":"crossref","first-page":"227","DOI":"10.1038\/s42256-022-00457-9","article-title":"Learning functional properties of proteins with language models","volume":"4","author":"Unsal","year":"2022","journal-title":"Nat Mach Intell"},{"key":"2026042409465017700_btag121-B25","doi-asserted-by":"crossref","first-page":"bbae177","DOI":"10.1093\/bib\/bbae177","article-title":"Evaluating large language models for annotating proteins","volume":"25","author":"Vitale","year":"2024","journal-title":"Brief Bioinform"},{"key":"2026042409465017700_btag121-B26","doi-asserted-by":"crossref","first-page":"337","DOI":"10.1089\/cmb.1994.1.337","article-title":"On the complexity of multiple sequence alignment","volume":"1","author":"Wand","year":"1994","journal-title":"J Comput Biol"},{"key":"2026042409465017700_btag121-B27","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/s40537-016-0043-6","article-title":"A survey of transfer learning","volume":"3","author":"Weiss","year":"2016","journal-title":"J Big Data"},{"key":"2026042409465017700_btag121-B28","doi-asserted-by":"crossref","first-page":"102997","DOI":"10.1016\/j.sbi.2025.102997","article-title":"Are protein language models the new universal key?","volume":"91","author":"Weissenow","year":"2025","journal-title":"Curr Opin Struct Biol"},{"key":"2026042409465017700_btag121-B29","doi-asserted-by":"crossref","first-page":"2642","DOI":"10.1093\/bioinformatics\/bty178","article-title":"Learned protein embeddings for machine learning","volume":"34","author":"Yang","year":"2018","journal-title":"Bioinformatics"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btag121\/67346247\/btag121.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/4\/btag121\/67346247\/btag121.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/4\/btag121\/67346247\/btag121.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,24]],"date-time":"2026-04-24T13:47:05Z","timestamp":1777038425000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/doi\/10.1093\/bioinformatics\/btag121\/8519617"}},"subtitle":[],"editor":[{"given":"Jianlin","family":"Cheng","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2026,3,13]]},"references-count":29,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,4,7]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btag121","relation":{},"ISSN":["1367-4803","1367-4811"],"issn-type":[{"value":"1367-4803","type":"print"},{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,4]]},"published":{"date-parts":[[2026,3,13]]},"article-number":"btag121"}}