{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,24]],"date-time":"2026-10-24T02:27:24Z","timestamp":1792808844524,"version":"4.5.0"},"reference-count":69,"publisher":"Wiley","issue":"2","license":[{"start":{"date-parts":[[2024,1,23]],"date-time":"2024-01-23T00:00:00Z","timestamp":1705968000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Protein Science"],"published-print":{"date-parts":[[2024,2]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Conventional protein\u2013protein docking algorithms usually rely on heavy candidate sampling and reranking, but these steps are time\u2010consuming and hinder applications that require high\u2010throughput complex structure prediction, for example, structure\u2010based virtual screening. Existing deep learning methods for protein\u2013protein docking, despite being much faster, suffer from low docking success rates. In addition, they simplify the problem to assume no conformational changes within any protein upon binding (rigid docking). This assumption precludes applications when binding\u2010induced conformational changes play a role, such as allosteric inhibition or docking from uncertain unbound model structures. To address these limitations, we present GeoDock, a multitrack iterative transformer network to predict a docked structure from separate docking partners. Unlike deep learning models for protein structure prediction that input multiple sequence alignments, GeoDock inputs just the sequences and structures of the docking partners, which suits the tasks when the individual structures are given. GeoDock is flexible at the protein residue level, allowing the prediction of conformational changes upon binding. On the Database of Interacting Protein Structures (DIPS) test set, GeoDock achieves a 43% top\u20101 success rate, outperforming all other tested methods. However, in the standard DIPS train\/test splits, we discovered contamination of close homologs in the training set. After decontaminating the training set, the success rate is 31%. On the DB5.5 test set and a benchmark dataset of antibody\u2013antigen complexes, GeoDock outperforms the deep learning models trained using the same dataset but falls behind most of the conventional methods and AlphaFold\u2010Multimer. GeoDock attains an average inference speed of under 1\u2009s on a single GPU, enabling its application in large\u2010scale structure screening. Although binding\u2010induced conformational changes are still a challenge owing to limited training and evaluation data, our architecture sets up the foundation to capture this backbone flexibility. Code and a demonstration Jupyter notebook are available at\n                    <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"https:\/\/github.com\/Graylab\/GeoDock\">https:\/\/github.com\/Graylab\/GeoDock<\/jats:ext-link>\n                    .\n                  <\/jats:p>","DOI":"10.1002\/pro.4862","type":"journal-article","created":{"date-parts":[[2023,12,26]],"date-time":"2023-12-26T21:45:08Z","timestamp":1703627108000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":35,"title":["Flexible protein\u2013protein docking with a multitrack iterative transformer"],"prefix":"10.1002","volume":"33","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8315-642X","authenticated-orcid":false,"given":"Lee\u2010Shin","family":"Chu","sequence":"first","affiliation":[{"name":"Department of Chemical and Biomolecular Engineering Johns Hopkins University Baltimore Maryland USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jeffrey A.","family":"Ruffolo","sequence":"additional","affiliation":[{"name":"Program in Molecular Biophysics Johns Hopkins University Baltimore Maryland USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ameya","family":"Harmalkar","sequence":"additional","affiliation":[{"name":"Department of Chemical and Biomolecular Engineering Johns Hopkins University Baltimore Maryland USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6380-2324","authenticated-orcid":false,"given":"Jeffrey J.","family":"Gray","sequence":"additional","affiliation":[{"name":"Department of Chemical and Biomolecular Engineering Johns Hopkins University Baltimore Maryland USA"},{"name":"Program in Molecular Biophysics Johns Hopkins University Baltimore Maryland USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2024,1,23]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.1002\/jcc.540150503"},{"key":"e_1_2_9_3_1","doi-asserted-by":"publisher","DOI":"10.1021\/acs.jctc.7b00125"},{"key":"e_1_2_9_4_1","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkh131"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1126\/science.abj8754"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.20770"},{"key":"e_1_2_9_7_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0161879"},{"key":"e_1_2_9_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jmb.2008.05.042"},{"key":"e_1_2_9_9_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.10389"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.10092"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkh354"},{"key":"e_1_2_9_12_1","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btab154"},{"key":"e_1_2_9_13_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.21723"},{"key":"e_1_2_9_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.str.2020.06.006"},{"key":"e_1_2_9_15_1","doi-asserted-by":"publisher","DOI":"10.1021\/ja026939x"},{"key":"e_1_2_9_16_1","doi-asserted-by":"publisher","DOI":"10.1007\/3-540-45784-4_14"},{"key":"e_1_2_9_17_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.24079"},{"key":"e_1_2_9_18_1","unstructured":"EvansR O'NeillM PritzelA AntropovaN SeniorAW GreenT et al.Protein complex prediction with AlphaFold\u2010Multimer. BioRxiv.2021."},{"key":"e_1_2_9_19_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.10383"},{"key":"e_1_2_9_20_1","first-page":"6533","article-title":"Protein interface prediction using graph convolutional networks","volume":"30","author":"Fout A","year":"2017","journal-title":"Adv Neural Inform Process Syst"},{"key":"e_1_2_9_21_1","doi-asserted-by":"publisher","DOI":"10.1006\/jmbi.1997.1203"},{"key":"e_1_2_9_22_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-019-0666-6"},{"key":"e_1_2_9_23_1","unstructured":"GaneaO\u2010E HuangX BunneC BianY BarzilayR JaakkolaT et al.Independent se (3)\u2010equivariant models for end\u2010to\u2010end rigid protein docking. arXiv preprint arXiv:2111.07786.2021."},{"key":"e_1_2_9_24_1","doi-asserted-by":"publisher","DOI":"10.1038\/ncomms100"},{"key":"e_1_2_9_25_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.1070"},{"key":"e_1_2_9_26_1","doi-asserted-by":"crossref","unstructured":"GhaniU DestaI JindalA KhanO JonesG HashemiN et al.Improved docking of protein models by a combination of alphafold2 and cluspro. BioRxiv 2021\u20139.2022.","DOI":"10.1101\/2021.09.07.459290"},{"key":"e_1_2_9_27_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0022-2836(03)00670-3"},{"key":"e_1_2_9_28_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.str.2021.01.005"},{"key":"e_1_2_9_29_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pcbi.1010124"},{"key":"e_1_2_9_30_1","doi-asserted-by":"publisher","DOI":"10.1110\/ps.26002"},{"key":"e_1_2_9_31_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.drudis.2014.02.005"},{"key":"e_1_2_9_32_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.10381"},{"key":"e_1_2_9_33_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-021-03819-2"},{"key":"e_1_2_9_34_1","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.89.6.2195"},{"key":"e_1_2_9_35_1","unstructured":"KetataMA LaueC MammadovR St\u00e4rkH WuM CorsoG et al.Diffdock\u2010pp: rigid protein\u2010protein docking with diffusion models. arXiv preprint arXiv:2304.03889.2023."},{"key":"e_1_2_9_36_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.21117"},{"key":"e_1_2_9_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/0022-2836(82)90153-X"},{"key":"e_1_2_9_38_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.str.2016.06.025"},{"key":"e_1_2_9_39_1","doi-asserted-by":"publisher","DOI":"10.1126\/science.ade2574"},{"key":"e_1_2_9_40_1","doi-asserted-by":"crossref","unstructured":"LiuY YuanH CaiL JiS.Deep learning of high\u2010order interactions for protein interface prediction. In Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining 679\u20136872020.","DOI":"10.1145\/3394486.3403110"},{"key":"e_1_2_9_41_1","doi-asserted-by":"publisher","DOI":"10.1093\/protein\/14.2.105"},{"key":"e_1_2_9_42_1","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btt473"},{"key":"e_1_2_9_43_1","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/bty355"},{"key":"e_1_2_9_44_1","doi-asserted-by":"crossref","unstructured":"McPartlonM XuJ.Deep learning for flexible and site\u2010specific protein docking and design. BioRxiv.2023.","DOI":"10.1101\/2023.04.01.535079"},{"key":"e_1_2_9_45_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-022-01488-1"},{"key":"e_1_2_9_46_1","doi-asserted-by":"publisher","DOI":"10.3390\/ijms11103623"},{"key":"e_1_2_9_47_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-021-91827-7"},{"key":"e_1_2_9_48_1","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2016239118"},{"key":"e_1_2_9_49_1","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btac063"},{"key":"e_1_2_9_50_1","unstructured":"RuidongW DingF WangR ShenR ZhangX LuoS et al.High\u2010resolution de novo structure prediction from primary sequence. BioRxiv.2022."},{"key":"e_1_2_9_51_1","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gki481"},{"key":"e_1_2_9_52_1","unstructured":"SoltanikazemiE RoyRS QuadirF GiriN MoreheadA ChengJ.Drlcomplex: reconstruction of protein quaternary structures using deep reinforcement learning. arXiv preprint arXiv:2205.13594.2022."},{"key":"e_1_2_9_53_1","unstructured":"SverrissonF FeydyJ CorreiaBE BronsteinMM.Fast end\u2010to\u2010end learning on protein surfaces. In Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition 15272\u2013152812021."},{"key":"e_1_2_9_54_1","unstructured":"SverrissonF FeydyJ SouthernJ BronsteinMM CorreiaB.Physics\u2010informed deep neural network for rigid\u2010body protein docking. In ICLR2022 machine learning for drug discovery.2022."},{"key":"e_1_2_9_55_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.sbi.2013.11.005"},{"key":"e_1_2_9_56_1","first-page":"15642","article-title":"End\u2010to\u2010end learning on 3d protein structure for interface prediction","volume":"32","author":"Townshend R","year":"2019","journal-title":"Adv Neural Inform Process Syst"},{"key":"e_1_2_9_57_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-022-01490-7"},{"key":"e_1_2_9_58_1","doi-asserted-by":"publisher","DOI":"10.1002\/(SICI)1097-0134(1997)1 <226::AID-PROT31>3.0.CO;2-O"},{"key":"e_1_2_9_59_1","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkab1061"},{"key":"e_1_2_9_60_1","doi-asserted-by":"publisher","DOI":"10.1186\/1471-2105-10-407"},{"key":"e_1_2_9_61_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jmb.2015.07.016"},{"key":"e_1_2_9_62_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jmb.2007.07.050"},{"key":"e_1_2_9_63_1","unstructured":"WangG FangX WuZ LiuY XueY XiangY et al.Helixfold: an efficient implementation of alphafold2 using paddlepaddle. arXiv preprint arXiv:2207.05477.2022."},{"key":"e_1_2_9_64_1","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btab761"},{"key":"e_1_2_9_65_1","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbab038"},{"key":"e_1_2_9_66_1","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1914677117"},{"key":"e_1_2_9_67_1","doi-asserted-by":"publisher","DOI":"10.1002\/pro.4379"},{"key":"e_1_2_9_68_1","doi-asserted-by":"crossref","unstructured":"YinR PierceBG.Evaluation of alphafold antibody\u2010antigen modeling with implications for improving predictive accuracy.Protein Sci.2023;33(1):e4865.","DOI":"10.1002\/pro.4865"},{"key":"e_1_2_9_69_1","doi-asserted-by":"publisher","DOI":"10.1110\/ps.0239303"},{"key":"e_1_2_9_70_1","doi-asserted-by":"publisher","DOI":"10.1002\/prot.20566"}],"container-title":["Protein Science"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/pro.4862","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,11,6]],"date-time":"2024-11-06T17:00:49Z","timestamp":1730912449000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/pro.4862"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,1,23]]},"references-count":69,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2024,2]]}},"alternative-id":["10.1002\/pro.4862"],"URL":"https:\/\/doi.org\/10.1002\/pro.4862","archive":["Portico"],"relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2023.06.29.547134","asserted-by":"object"}]},"ISSN":["0961-8368","1469-896X"],"issn-type":[{"value":"0961-8368","type":"print"},{"value":"1469-896X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,1,23]]},"assertion":[{"value":"2023-07-09","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-12-06","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-01-23","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e4862"}}