{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,19]],"date-time":"2026-06-19T15:21:06Z","timestamp":1781882466384,"version":"3.54.5"},"reference-count":63,"publisher":"Oxford University Press (OUP)","issue":"4","license":[{"start":{"date-parts":[[2023,6,16]],"date-time":"2023-06-16T00:00:00Z","timestamp":1686873600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/pages\/standard-publication-reuse-rights"}],"funder":[{"name":"Technology and Innovation Major Project of the Ministry of Science and Technology of China","award":["2020AAA0108400"],"award-info":[{"award-number":["2020AAA0108400"]}]},{"name":"Technology and Innovation Major Project of the Ministry of Science and Technology of China","award":["2020AAA0108402"],"award-info":[{"award-number":["2020AAA0108402"]}]},{"name":"NSFC for Distinguished Young Scholar","award":["61825602"],"award-info":[{"award-number":["61825602"]}]},{"name":"Zhejiang Lab","award":["2022PE0AC04"],"award-info":[{"award-number":["2022PE0AC04"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2023,7,20]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>AlphaFold-Multimer has greatly improved the protein complex structure prediction, but its accuracy also depends on the quality of the multiple sequence alignment (MSA) formed by the interacting homologs (i.e. interologs) of the complex under prediction. Here we propose a novel method, ESMPair, that can identify interologs of a complex using protein language models. We show that ESMPair can generate better interologs than the default MSA generation method in AlphaFold-Multimer. Our method results in better complex structure prediction than AlphaFold-Multimer by a large margin (+10.7% in terms of the Top-5 best DockQ), especially when the predicted complex structures have low confidence. We further show that by combining several MSA generation methods, we may yield even better complex structure prediction accuracy than Alphafold-Multimer (+22% in terms of the Top-5 best DockQ). By systematically analyzing the impact factors of our algorithm we find that the diversity of MSA of interologs significantly affects the prediction accuracy. Moreover, we show that ESMPair performs particularly well on complexes in eucaryotes.<\/jats:p>","DOI":"10.1093\/bib\/bbad221","type":"journal-article","created":{"date-parts":[[2023,6,17]],"date-time":"2023-06-17T03:06:24Z","timestamp":1686971184000},"source":"Crossref","is-referenced-by-count":25,"title":["Improved the heterodimer protein complex prediction with protein language models"],"prefix":"10.1093","volume":"24","author":[{"given":"Bo","family":"Chen","sequence":"first","affiliation":[{"name":"Department of Computer Science and Technology, Tsinghua University , Beijing , China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ziwei","family":"Xie","sequence":"additional","affiliation":[{"name":"Toyota Technological Institute at Chicago , Chicago, IL 60637 , USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiezhong","family":"Qiu","sequence":"additional","affiliation":[{"name":"Zhejiang Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhaofeng","family":"Ye","sequence":"additional","affiliation":[{"name":"MOE Key Laboratory of Bioinformatics , School of Medicine, , Beijing 100084 , China"},{"name":"Tsinghua University , School of Medicine, , Beijing 100084 , China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jinbo","family":"Xu","sequence":"additional","affiliation":[{"name":"Toyota Technological Institute at Chicago , Chicago, IL 60637 , USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jie","family":"Tang","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, Tsinghua University , Beijing , China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2023,6,16]]},"reference":[{"issue":"1","key":"2023072020155981100_ref1","doi-asserted-by":"crossref","first-page":"13","DOI":"10.1073\/pnas.93.1.13","article-title":"Thornton. Principles of protein\u2013protein interactions","volume":"93","author":"Jones","year":"1996","journal-title":"Proc Natl Acad Sci"},{"key":"2023072020155981100_ref2","first-page":"3","article-title":"Structural basis of protein-protein interactions. Protein-protein","author":"Liddington","year":"2004","journal-title":"Interactions"},{"issue":"6","key":"2023072020155981100_ref3","doi-asserted-by":"crossref","first-page":"1974","DOI":"10.1073\/pnas.0409522102","article-title":"Conserved patterns of protein interaction in multiple species","volume":"102","author":"Sharan","year":"2005","journal-title":"Proc Natl Acad Sci"},{"issue":"2","key":"2023072020155981100_ref4","doi-asserted-by":"crossref","first-page":"67","DOI":"10.1038\/gene.2012.55","article-title":"Common and specific signatures of gene expression and protein\u2013protein interactions in autoimmune diseases","volume":"14","author":"Tuller","year":"2013","journal-title":"Genes Immun"},{"issue":"6295","key":"2023072020155981100_ref5","doi-asserted-by":"crossref","first-page":"123","DOI":"10.1126\/science.aah3449","article-title":"Network analytics in the age of big data","volume":"353","author":"Pr\u017eulj","year":"2016","journal-title":"Science"},{"issue":"2","key":"2023072020155981100_ref6","doi-asserted-by":"crossref","first-page":"255","DOI":"10.1038\/nprot.2016.169","article-title":"The cluspro web server for protein\u2013protein docking","volume":"12","author":"Kozakov","year":"2017","journal-title":"Nat Protoc"},{"issue":"2","key":"2023072020155981100_ref7","doi-asserted-by":"crossref","first-page":"392","DOI":"10.1002\/prot.21117","article-title":"Piper: an fft-based protein docking program with pairwise potentials","volume":"65","author":"Kozakov","year":"2006","journal-title":"Proteins"},{"issue":"12","key":"2023072020155981100_ref8","doi-asserted-by":"crossref","first-page":"1771","DOI":"10.1093\/bioinformatics\/btu097","article-title":"Zdock server: interactive docking prediction of protein\u2013protein complexes and symmetric multimers","volume":"30","author":"Pierce","year":"2014","journal-title":"Bioinformatics"},{"issue":"suppl_2","key":"2023072020155981100_ref9","doi-asserted-by":"crossref","first-page":"W233","DOI":"10.1093\/nar\/gkn216","article-title":"The rosettadock server for local protein\u2013protein docking","volume":"36","author":"Lyskov","year":"2008","journal-title":"Nucleic Acids Res"},{"issue":"9","key":"2023072020155981100_ref10","doi-asserted-by":"crossref","first-page":"1071","DOI":"10.1016\/j.str.2020.06.006","article-title":"Performance and its limits in rigid body protein-protein docking","volume":"28","author":"Desta","year":"2020","journal-title":"Structure"},{"issue":"4","key":"2023072020155981100_ref11","doi-asserted-by":"crossref","first-page":"1225","DOI":"10.1021\/cr040409x","article-title":"Principles of protein- protein interactions: what are the preferred ways for proteins to interact?","volume":"108","author":"Keskin","year":"2008","journal-title":"Chem Rev"},{"issue":"14","key":"2023072020155981100_ref12","doi-asserted-by":"crossref","first-page":"3486","DOI":"10.1093\/emboj\/cdg359","article-title":"Diversity of protein\u2013protein interactions","volume":"22","author":"Nooren","year":"2003","journal-title":"EMBO J"},{"issue":"1","key":"2023072020155981100_ref13","doi-asserted-by":"crossref","first-page":"80","DOI":"10.1002\/prot.10389","article-title":"Zdock: an initial-stage protein-docking algorithm","volume":"52","author":"Chen","year":"2003","journal-title":"Proteins"},{"issue":"7","key":"2023072020155981100_ref14","doi-asserted-by":"crossref","first-page":"1731","DOI":"10.1021\/ja026939x","article-title":"Haddock: a protein- protein docking approach based on biochemical or biophysical information","volume":"125","author":"Dominguez","year":"2003","journal-title":"J Am Chem Soc"},{"issue":"8","key":"2023072020155981100_ref15","doi-asserted-by":"crossref","first-page":"e4379","DOI":"10.1002\/pro.4379","article-title":"Benchmarking alphafold for protein complex modeling reveals accuracy determinants","volume":"31","author":"Yin","year":"2022","journal-title":"Protein Sci"},{"issue":"1","key":"2023072020155981100_ref16","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-021-87524-0","article-title":"The whole is greater than its parts: ensembling improves protein contact prediction","volume":"11","author":"Billings","year":"2021","journal-title":"Sci Rep"},{"issue":"7","key":"2023072020155981100_ref17","doi-asserted-by":"crossref","first-page":"1888","DOI":"10.1093\/bioinformatics\/btac053","article-title":"Spot-contact-lm: improving single-sequence-based prediction of protein contact map using a transformer language model","volume":"38","author":"Singh","year":"2022","journal-title":"Bioinformatics"},{"key":"2023072020155981100_ref18","article-title":"Co-evolution transformer for protein contact prediction","volume":"34","author":"Zhang","year":"2021","journal-title":"Adv Neural Inf Process Syst"},{"issue":"7873","key":"2023072020155981100_ref19","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"},{"issue":"6557","key":"2023072020155981100_ref20","doi-asserted-by":"crossref","first-page":"871","DOI":"10.1126\/science.abj8754","article-title":"Accurate prediction of protein structures and interactions using a three-track neural network","volume":"373","author":"Baek","year":"2021","journal-title":"Science"},{"issue":"7","key":"2023072020155981100_ref21","doi-asserted-by":"crossref","first-page":"1904","DOI":"10.1093\/bioinformatics\/btac063","article-title":"A deep dilated convolutional residual network for predicting interchain contacts of protein homodimers","volume":"38","author":"Roy","year":"2022","journal-title":"Bioinformatics"},{"issue":"1","key":"2023072020155981100_ref22","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-020-60598-y","article-title":"Deep learning to predict protein backbone structure from high-resolution cryo-em density maps","volume":"10","author":"Si","year":"2020","journal-title":"Sci Rep"},{"issue":"1","key":"2023072020155981100_ref23","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s42003-021-02399-1","article-title":"Deepemhancer: a deep learning solution for cryo-em volume post-processing","volume":"4","author":"Sanchez-Garcia","year":"2021","journal-title":"Commun Biol"},{"key":"2023072020155981100_ref24","first-page":"240754","article-title":"Deep learning reveals many more inter-protein residue-residue contacts than direct coupling analysis","author":"Zhou","year":"2018","journal-title":"bioRxiv"},{"issue":"4","key":"2023072020155981100_ref25","doi-asserted-by":"crossref","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"},{"issue":"1","key":"2023072020155981100_ref26","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/ncomms100","article-title":"Af2complex predicts direct physical interactions in multimeric proteins with deep learning","volume":"13","author":"Gao","year":"2022","journal-title":"Nat Commun"},{"issue":"1","key":"2023072020155981100_ref27","first-page":"1","article-title":"Improved prediction of protein-protein interactions using alphafold2","volume":"13","author":"Bryant","year":"2022","journal-title":"Nat Commun"},{"key":"2023072020155981100_ref28","article-title":"Protein complex prediction with alphafold-multimer","author":"Evans","year":"2021","journal-title":"BioRxiv"},{"issue":"8","key":"2023072020155981100_ref29","doi-asserted-by":"crossref","first-page":"4285","DOI":"10.1073\/pnas.96.8.4285","article-title":"Assigning protein functions by comparative genome analysis: protein phylogenetic profiles","volume":"96","author":"Pellegrini","year":"1999","journal-title":"Proc Natl Acad Sci"},{"issue":"3","key":"2023072020155981100_ref30","doi-asserted-by":"crossref","first-page":"934","DOI":"10.1073\/pnas.0709671105","article-title":"High-confidence prediction of global interactomes based on genome-wide coevolutionary networks","volume":"105","author":"Juan","year":"2008","journal-title":"Proc Natl Acad Sci"},{"issue":"W1","key":"2023072020155981100_ref31","doi-asserted-by":"crossref","first-page":"W432","DOI":"10.1093\/nar\/gky420","article-title":"Complexcontact: a web server for inter-protein contact prediction using deep learning","volume":"46","author":"Zeng","year":"2018","journal-title":"Nucleic Acids Res"},{"issue":"2","key":"2023072020155981100_ref32","doi-asserted-by":"crossref","first-page":"e0149166","DOI":"10.1371\/journal.pone.0149166","article-title":"Inter-protein sequence co-evolution predicts known physical interactions in bacterial ribosomes and the trp operon","volume":"11","author":"Feinauer","year":"2016","journal-title":"PloS One"},{"key":"2023072020155981100_ref33","doi-asserted-by":"crossref","first-page":"e02030","DOI":"10.7554\/eLife.02030","article-title":"Robust and accurate prediction of residue\u2013residue interactions across protein interfaces using evolutionary information","volume":"3","author":"Ovchinnikov","year":"2014","journal-title":"Elife"},{"issue":"52","key":"2023072020155981100_ref34","doi-asserted-by":"crossref","first-page":"15018","DOI":"10.1073\/pnas.1611861114","article-title":"Conservation of coevolving protein interfaces bridges prokaryote\u2013eukaryote homologies in the twilight zone","volume":"113","author":"Rodriguez-Rivas","year":"2016","journal-title":"Proc Natl Acad Sci"},{"issue":"suppl_1","key":"2023072020155981100_ref35","doi-asserted-by":"crossref","first-page":"D115","DOI":"10.1093\/nar\/gkh131","article-title":"Uniprot: the universal protein knowledgebase","volume":"32","author":"Apweiler","year":"2004","journal-title":"Nucleic Acids Res"},{"issue":"1","key":"2023072020155981100_ref36","first-page":"1","article-title":"Hidden markov model speed heuristic and iterative hmm search procedure","volume":"11","author":"Steven Johnson","year":"2010","journal-title":"BMC Bioinformatics"},{"key":"2023072020155981100_ref37","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","author":"Brown","year":"2020","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023072020155981100_ref38","doi-asserted-by":"crossref","first-page":"1150","DOI":"10.1145\/3394486.3403168","article-title":"Gcc: Graph contrastive coding for graph neural network pre-training","volume-title":"Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining","author":"Qiu","year":"2020"},{"key":"2023072020155981100_ref39","first-page":"4171","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","volume-title":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Devlin","year":"2019"},{"key":"2023072020155981100_ref40","article-title":"An image is worth 16x16 words: Transformers for image recognition at scale","volume-title":"International Conference on Learning Representations","author":"Dosovitskiy","year":"2020"},{"key":"2023072020155981100_ref41","first-page":"2020","article-title":"ProtTrans: towards cracking the language of lifes code through self-supervised deep learning and high performance computing","author":"Elnaggar","year":"2021","journal-title":"IEEE transactions on pattern analysis and machine intelligence"},{"key":"2023072020155981100_ref42","first-page":"8844","article-title":"Msa transformer","volume-title":"International Conference on Machine Learning","author":"Rao","year":"2021"},{"issue":"15","key":"2023072020155981100_ref43","doi-asserted-by":"crossref","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"},{"key":"2023072020155981100_ref44","article-title":"Evaluating protein transfer learning with tape","volume":"32","author":"Rao","year":"2019","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023072020155981100_ref45","doi-asserted-by":"crossref","DOI":"10.1101\/2020.06.26.174417","article-title":"Bertology meets biology: Interpreting attention in protein language models","volume-title":"International Conference on Learning Representations","author":"Vig","year":"2020"},{"key":"2023072020155981100_ref46","article-title":"Language models enable zero-shot prediction of the effects of mutations on protein function","volume":"34","author":"Meier","year":"2021","journal-title":"Adv Neural Inf Process Syst"},{"issue":"10","key":"2023072020155981100_ref47","first-page":"577","article-title":"Neurofilaments as biomarkers in neurological disorders. Nature reviews","volume":"14","author":"Khalil","year":"2018","journal-title":"Neurology"},{"issue":"8","key":"2023072020155981100_ref48","doi-asserted-by":"crossref","first-page":"eaax5783","DOI":"10.1126\/sciadv.aax5783","article-title":"Stitching the synapse: cross-linking mass spectrometry into resolving synaptic protein interactions","volume":"6","author":"Gonzalez-Lozano","year":"2020","journal-title":"Sci Adv"},{"issue":"4","key":"2023072020155981100_ref49","doi-asserted-by":"crossref","DOI":"10.1093\/bib\/bbac208","article-title":"Protein complex structure prediction powered by multiple sequence alignment of interologs from multiple taxonomic ranks and alphafold2","volume":"23","author":"Si","year":"2022","journal-title":"Brief Bioinform"},{"key":"2023072020155981100_ref50","first-page":"4947","article-title":"How transfer learning impacts linguistic knowledge in deep nlp models?","volume":"2021","author":"Durrani","year":"2021","journal-title":"Findings of the Association for Computational Linguistics: ACL-IJCNLP"},{"key":"2023072020155981100_ref51","doi-asserted-by":"crossref","first-page":"33","DOI":"10.18653\/v1\/2020.blackboxnlp-1.4","article-title":"What happens to bert embeddings during fine-tuning?","author":"Merchant","year":"2020","journal-title":"Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP"},{"key":"2023072020155981100_ref52","doi-asserted-by":"crossref","first-page":"375","DOI":"10.18653\/v1\/2021.blackboxnlp-1.29","article-title":"Not all models localize linguistic knowledge in the same place: A layer-wise probing on bertoids\u2019 representations","volume-title":"Proceedings of the Fourth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP","author":"Fayyaz","year":"2021"},{"key":"2023072020155981100_ref53","article-title":"Axial attention in multidimensional transformers","author":"Ho","year":"2019"},{"key":"2023072020155981100_ref54","first-page":"603","article-title":"Ccnet: Criss-cross attention for semantic segmentation","volume-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision","author":"Huang","year":"2019"},{"key":"2023072020155981100_ref55","first-page":"1597","article-title":"A simple framework for contrastive learning of visual representations","volume-title":"International conference on machine learning","author":"Chen","year":"2020"},{"key":"2023072020155981100_ref56","doi-asserted-by":"crossref","first-page":"6894","DOI":"10.18653\/v1\/2021.emnlp-main.552","article-title":"SimCSE: Simple contrastive learning of sentence embeddings","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Gao","year":"2021"},{"issue":"1","key":"2023072020155981100_ref57","doi-asserted-by":"crossref","first-page":"32","DOI":"10.1137\/0105003","article-title":"Algorithms for the assignment and transportation problems","volume":"5","author":"Munkres","year":"1957","journal-title":"J Soc Ind Appl Math"},{"issue":"8","key":"2023072020155981100_ref58","doi-asserted-by":"crossref","first-page":"e0161879","DOI":"10.1371\/journal.pone.0161879","article-title":"Dockq: a quality measure for protein-protein docking models","volume":"11","author":"Basu","year":"2016","journal-title":"PloS One"},{"issue":"D1","key":"2023072020155981100_ref59","doi-asserted-by":"crossref","first-page":"D170","DOI":"10.1093\/nar\/gkw1081","article-title":"Uniclust databases of clustered and deeply annotated protein sequences and alignments","volume":"45","author":"Mirdita","year":"2017","journal-title":"Nucleic Acids Res"},{"issue":"2","key":"2023072020155981100_ref60","doi-asserted-by":"crossref","first-page":"173","DOI":"10.1038\/nmeth.1818","article-title":"Hhblits: lightning-fast iterative protein sequence searching by hmm-hmm alignment","volume":"9","author":"Remmert","year":"2012","journal-title":"Nat Methods"},{"issue":"43","key":"2023072020155981100_ref61","doi-asserted-by":"crossref","first-page":"12186","DOI":"10.1073\/pnas.1607570113","article-title":"Simultaneous identification of specifically interacting paralogs and interprotein contacts by direct coupling analysis","volume":"113","author":"Gueudr\u00e9","year":"2016","journal-title":"Proc Natl Acad Sci"},{"issue":"1","key":"2023072020155981100_ref62","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/s12859-019-3190-x","article-title":"Deepeca: an end-to-end learning framework for protein contact prediction from a multiple sequence alignment","volume":"21","author":"Fukuda","year":"2020","journal-title":"BMC Bioinformatics"},{"issue":"2","key":"2023072020155981100_ref63","doi-asserted-by":"crossref","first-page":"e0169356","DOI":"10.1371\/journal.pone.0169356","article-title":"Improving protein-protein interaction prediction using evolutionary information from low-quality msas","volume":"12","author":"Varnai","year":"2017","journal-title":"PloS One"}],"container-title":["Briefings in Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/24\/4\/bbad221\/50917338\/bbad221.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/24\/4\/bbad221\/50917338\/bbad221.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,7,20]],"date-time":"2023-07-20T20:19:14Z","timestamp":1689884354000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bib\/article\/doi\/10.1093\/bib\/bbad221\/7199490"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,6,16]]},"references-count":63,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2023,7,20]]}},"URL":"https:\/\/doi.org\/10.1093\/bib\/bbad221","relation":{},"ISSN":["1467-5463","1477-4054"],"issn-type":[{"value":"1467-5463","type":"print"},{"value":"1477-4054","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2023,7]]},"published":{"date-parts":[[2023,6,16]]},"article-number":"bbad221"}}