{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,10]],"date-time":"2026-04-10T02:11:22Z","timestamp":1775787082780,"version":"3.50.1"},"reference-count":50,"publisher":"Oxford University Press (OUP)","issue":"20","license":[{"start":{"date-parts":[[2020,7,17]],"date-time":"2020-07-17T00:00:00Z","timestamp":1594944000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/journals\/pages\/open_access\/funder_policies\/chorus\/standard_publication_model"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Basic Research Program of China","doi-asserted-by":"publisher","award":["2019YFC1711600"],"award-info":[{"award-number":["2019YFC1711600"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Shanghai Municipal Science and Technology Major","award":["2018SHZDZX01"],"award-info":[{"award-number":["2018SHZDZX01"]}]},{"DOI":"10.13039\/100000928","name":"Welch Foundation","doi-asserted-by":"publisher","award":["Q-1826"],"award-info":[{"award-number":["Q-1826"]}],"id":[{"id":"10.13039\/100000928","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000928","name":"Welch Foundation","doi-asserted-by":"publisher","award":["Q-1512"],"award-info":[{"award-number":["Q-1512"]}],"id":[{"id":"10.13039\/100000928","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2020,12,22]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Motivation<\/jats:title>\n                  <jats:p>Predictions of protein backbone torsion angles (\u03d5 and \u03c8) and secondary structure from sequence are crucial subproblems in protein structure prediction. With the development of deep learning approaches, their accuracies have been significantly improved. To capture the long-range interactions, most studies integrate bidirectional recurrent neural networks into their models. In this study, we introduce and modify a recently proposed architecture named Transformer to capture the interactions between the two residues theoretically with arbitrary distance. Moreover, we take advantage of multitask learning to improve the generalization of neural network by introducing related tasks into the training process. Similar to many previous studies, OPUS-TASS uses an ensemble of models and achieves better results.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>OPUS-TASS uses the same training and validation sets as SPOT-1D. We compare the performance of OPUS-TASS and SPOT-1D on TEST2016 (1213 proteins) and TEST2018 (250 proteins) proposed in the SPOT-1D paper, CASP12 (55 proteins), CASP13 (32 proteins) and CASP-FM (56 proteins) proposed in the SAINT paper, and a recently released PDB structure collection from CAMEO (93 proteins) named as CAMEO93. On these six test sets, OPUS-TASS achieves consistent improvements in both backbone torsion angles prediction and secondary structure prediction. On CAMEO93, SPOT-1D achieves the mean absolute errors of 16.89 and 23.02 for \u03d5 and \u03c8 predictions, respectively, and the accuracies for 3- and 8-state secondary structure predictions are 87.72 and 77.15%, respectively. In comparison, OPUS-TASS achieves 16.56 and 22.56 for \u03d5 and \u03c8 predictions, and 89.06 and 78.87% for 3- and 8-state secondary structure predictions, respectively. In particular, after using our torsion angles refinement method OPUS-Refine as the post-processing procedure for OPUS-TASS, the mean absolute errors for final \u03d5 and \u03c8 predictions are further decreased to 16.28 and 21.98, respectively.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Availability and implementation<\/jats:title>\n                  <jats:p>The training and the inference codes of OPUS-TASS and its data are available at https:\/\/github.com\/thuxugang\/opus_tass.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Supplementary information<\/jats:title>\n                  <jats:p>Supplementary data are available at Bioinformatics online.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btaa629","type":"journal-article","created":{"date-parts":[[2020,7,10]],"date-time":"2020-07-10T11:34:24Z","timestamp":1594380864000},"page":"5021-5026","source":"Crossref","is-referenced-by-count":57,"title":["OPUS-TASS: a protein backbone torsion angles and secondary structure predictor based on ensemble neural networks"],"prefix":"10.1093","volume":"36","author":[{"given":"Gang","family":"Xu","sequence":"first","affiliation":[{"name":"Multiscale Research Institute of Complex Systems, Fudan University , Shanghai 200433, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qinghua","family":"Wang","sequence":"additional","affiliation":[{"name":"Verna and Marrs Mclean Department of Biochemistry and Molecular Biology, Baylor College of Medicine , One Baylor Plaza, Houston, TX 77030, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jianpeng","family":"Ma","sequence":"additional","affiliation":[{"name":"Multiscale Research Institute of Complex Systems, Fudan University , Shanghai 200433, China"},{"name":"Verna and Marrs Mclean Department of Biochemistry and Molecular Biology, Baylor College of Medicine , One Baylor Plaza, Houston, TX 77030, USA"},{"name":"Department of Bioengineering, Rice University , Houston, TX 77030, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2020,7,17]]},"reference":[{"key":"2023062408113504800_btaa629-B1","first-page":"265","author":"Abadi","year":"2016"},{"key":"2023062408113504800_btaa629-B2","doi-asserted-by":"crossref","first-page":"292","DOI":"10.1016\/j.cels.2019.03.006","article-title":"End-to-end differentiable learning of protein structure","volume":"8","author":"AlQuraishi","year":"2019","journal-title":"Cell Syst"},{"key":"2023062408113504800_btaa629-B3","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":"2023062408113504800_btaa629-B4","doi-asserted-by":"crossref","first-page":"592","DOI":"10.1002\/prot.25487","article-title":"MUFOLD-SS: new deep inception-inside-inception networks for protein secondary structure prediction","volume":"86","author":"Fang","year":"2018","journal-title":"Proteins"},{"key":"2023062408113504800_btaa629-B5","author":"Fang","year":"2018"},{"key":"2023062408113504800_btaa629-B6","doi-asserted-by":"crossref","first-page":"100","DOI":"10.1186\/s12859-018-2065-x","article-title":"RaptorX-Angle: real-value prediction of protein backbone dihedral angles through a hybrid method of clustering and deep learning","volume":"19","author":"Gao","year":"2018","journal-title":"BMC Bioinformatics"},{"key":"2023062408113504800_btaa629-B7","article-title":"Deep learning extends de novo protein modelling coverage of genomes using iteratively predicted structural constraints","volume":"10, 1-13","author":"Greener","year":"2019","journal-title":"Nat. Commun"},{"key":"2023062408113504800_btaa629-B8","doi-asserted-by":"crossref","first-page":"bat031","DOI":"10.1093\/database\/bat031","article-title":"The Protein Model Portal\u2014a comprehensive resource for protein structure and model information","volume":"2013","author":"Haas","year":"2013","journal-title":"Database (Oxford)"},{"key":"2023062408113504800_btaa629-B9","doi-asserted-by":"crossref","first-page":"4039","DOI":"10.1093\/bioinformatics\/bty481","article-title":"Accurate prediction of protein contact maps by coupling residual two-dimensional bidirectional long short-term memory with convolutional neural networks","volume":"34","author":"Hanson","year":"2018","journal-title":"Bioinformatics"},{"key":"2023062408113504800_btaa629-B10","doi-asserted-by":"crossref","first-page":"2403","DOI":"10.1093\/bioinformatics\/bty1006","article-title":"Improving prediction of protein secondary structure, backbone angles, solvent accessibility and contact numbers by using predicted contact maps and an ensemble of recurrent and residual convolutional neural networks","volume":"35","author":"Hanson","year":"2019","journal-title":"Bioinformatics"},{"key":"2023062408113504800_btaa629-B11","doi-asserted-by":"crossref","first-page":"11476","DOI":"10.1038\/srep11476","article-title":"Improving prediction of secondary structure, local backbone angles, and solvent accessible surface area of proteins by iterative deep learning","volume":"5","author":"Heffernan","year":"2015","journal-title":"Sci. Rep"},{"key":"2023062408113504800_btaa629-B12","doi-asserted-by":"crossref","first-page":"2842","DOI":"10.1093\/bioinformatics\/btx218","article-title":"Capturing non-local interactions by long short-term memory bidirectional recurrent neural networks for improving prediction of protein secondary structure, backbone angles, contact numbers and solvent accessibility","volume":"33","author":"Heffernan","year":"2017","journal-title":"Bioinformatics"},{"key":"2023062408113504800_btaa629-B13","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Comput"},{"key":"2023062408113504800_btaa629-B14","doi-asserted-by":"crossref","first-page":"1165","DOI":"10.1002\/prot.25697","article-title":"Protein tertiary structure modeling driven by deep learning and contact distance prediction in CASP13","volume":"87","author":"Hou","year":"2019","journal-title":"Proteins"},{"key":"2023062408113504800_btaa629-B15","first-page":"448","author":"Ioffe","year":"2015"},{"key":"2023062408113504800_btaa629-B16","doi-asserted-by":"crossref","first-page":"2577","DOI":"10.1002\/bip.360221211","article-title":"Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features","volume":"22","author":"Kabsch","year":"1983","journal-title":"Biopolymers"},{"key":"2023062408113504800_btaa629-B17","doi-asserted-by":"crossref","first-page":"W526","DOI":"10.1093\/nar\/gkh468","article-title":"Protein structure prediction and analysis using the Robetta server","volume":"32","author":"Kim","year":"2004","journal-title":"Nucleic Acids Res"},{"key":"2023062408113504800_btaa629-B18","author":"Kingma","year":"2015"},{"key":"2023062408113504800_btaa629-B19","doi-asserted-by":"crossref","first-page":"520","DOI":"10.1002\/prot.25674","article-title":"NetSurfP-2.0: improved prediction of protein structural features by integrated deep learning","volume":"87","author":"Klausen","year":"2019","journal-title":"Proteins"},{"key":"2023062408113504800_btaa629-B20","doi-asserted-by":"crossref","first-page":"436","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"2023062408113504800_btaa629-B21","doi-asserted-by":"crossref","first-page":"2278","DOI":"10.1109\/5.726791","article-title":"Gradient-based learning applied to document recognition","volume":"86","author":"Lecun","year":"1998","journal-title":"Proc. IEEE"},{"key":"2023062408113504800_btaa629-B22","author":"Lounici","year":"2009"},{"key":"2023062408113504800_btaa629-B23","doi-asserted-by":"crossref","first-page":"288","DOI":"10.1016\/j.jmb.2007.11.033","article-title":"OPUS-PSP: an orientation-dependent statistical all-atom potential derived from side-chain packing","volume":"376","author":"Lu","year":"2008","journal-title":"J. Mol. Biol"},{"key":"2023062408113504800_btaa629-B24","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"},{"key":"2023062408113504800_btaa629-B25","first-page":"9689","article-title":"Evaluating protein transfer learning with TAPE","author":"Rao","year":"2019"},{"key":"2023062408113504800_btaa629-B26","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"},{"key":"2023062408113504800_btaa629-B27","doi-asserted-by":"crossref","first-page":"204","DOI":"10.1006\/jsbi.2001.4336","article-title":"Review: protein secondary structure prediction continues to rise","volume":"134","author":"Rost","year":"2001","journal-title":"J. Struct. Biol"},{"key":"2023062408113504800_btaa629-B28","doi-asserted-by":"crossref","first-page":"2673","DOI":"10.1109\/78.650093","article-title":"Bidirectional recurrent neural networks","volume":"45","author":"Schuster","year":"1997","journal-title":"IEEE Trans. Signal Process"},{"key":"2023062408113504800_btaa629-B29","doi-asserted-by":"crossref","first-page":"706","DOI":"10.1038\/s41586-019-1923-7","article-title":"Improved protein structure prediction using potentials from deep learning","volume":"577","author":"Senior","year":"2020","journal-title":"Nature"},{"key":"2023062408113504800_btaa629-B30","doi-asserted-by":"crossref","first-page":"951","DOI":"10.1093\/bioinformatics\/bti125","article-title":"Protein homology detection by HMM-HMM comparison","volume":"21","author":"Soding","year":"2005","journal-title":"Bioinformatics"},{"key":"2023062408113504800_btaa629-B31","doi-asserted-by":"crossref","first-page":"e30361","DOI":"10.1371\/journal.pone.0030361","article-title":"TANGLE: two-level support vector regression approach for protein backbone torsion angle prediction from primary sequences","volume":"7","author":"Song","year":"2012","journal-title":"PLoS One"},{"key":"2023062408113504800_btaa629-B32","first-page":"1929","article-title":"Dropout: a simple way to prevent neural networks from overfitting","volume":"15","author":"Srivastava","year":"2014","journal-title":"J. Mach. Learn. Res"},{"key":"2023062408113504800_btaa629-B33","doi-asserted-by":"crossref","first-page":"473","DOI":"10.1186\/s12859-019-3019-7","article-title":"HH-suite3 for fast remote homology detection and deep protein annotation","volume":"20","author":"Steinegger","year":"2019","journal-title":"BMC Bioinformatics"},{"key":"2023062408113504800_btaa629-B34","doi-asserted-by":"crossref","first-page":"926","DOI":"10.1093\/bioinformatics\/btu739","article-title":"UniRef clusters: a comprehensive and scalable alternative for improving sequence similarity searches","volume":"31","author":"Suzek","year":"2015","journal-title":"Bioinformatics"},{"key":"2023062408113504800_btaa629-B35","first-page":"4278","author":"Szegedy","year":"2017"},{"key":"2023062408113504800_btaa629-B36","doi-asserted-by":"crossref","first-page":"12374","DOI":"10.1038\/s41598-019-48786-x","article-title":"Deeper profiles and cascaded recurrent and convolutional neural networks for state-of-the-art protein secondary structure prediction","volume":"9","author":"Torrisi","year":"2019","journal-title":"Sci. Rep"},{"key":"2023062408113504800_btaa629-B37","doi-asserted-by":"crossref","first-page":"D364","DOI":"10.1093\/nar\/gku1028","article-title":"A series of PDB-related databanks for everyday needs","volume":"43","author":"Touw","year":"2015","journal-title":"Nucleic Acids Res"},{"key":"2023062408113504800_btaa629-B38","doi-asserted-by":"crossref","DOI":"10.1093\/bioinformatics\/btaa531","article-title":"SAINT: self-attention augmented inception-inside-inception network improves protein secondary structure prediction","author":"Uddin","year":"2020","journal-title":"Bioinformatics"},{"key":"2023062408113504800_btaa629-B39","first-page":"5998","article-title":"Attention is all you need","author":"Vaswani","year":"2017"},{"key":"2023062408113504800_btaa629-B40","doi-asserted-by":"crossref","first-page":"1589","DOI":"10.1093\/bioinformatics\/btg224","article-title":"PISCES: a protein sequence culling server","volume":"19","author":"Wang","year":"2003","journal-title":"Bioinformatics"},{"key":"2023062408113504800_btaa629-B41","doi-asserted-by":"crossref","first-page":"W430","DOI":"10.1093\/nar\/gkw306","article-title":"RaptorX-Property: a web server for protein structure property prediction","volume":"44","author":"Wang","year":"2016","journal-title":"Nucleic Acids Res"},{"key":"2023062408113504800_btaa629-B42","doi-asserted-by":"crossref","first-page":"W296","DOI":"10.1093\/nar\/gky427","article-title":"SWISS-MODEL: homology modelling of protein structures and complexes","volume":"46","author":"Waterhouse","year":"2018","journal-title":"Nucleic Acids Res"},{"key":"2023062408113504800_btaa629-B43","doi-asserted-by":"crossref","first-page":"5.6.1","DOI":"10.1002\/cpbi.3","article-title":"Comparative protein structure modeling using MODELLER","volume":"54","author":"Webb","year":"2016","journal-title":"Curr. Protoc. Bioinf"},{"key":"2023062408113504800_btaa629-B44","doi-asserted-by":"crossref","first-page":"e3400","DOI":"10.1371\/journal.pone.0003400","article-title":"ANGLOR: a composite machine-learning algorithm for protein backbone torsion angle prediction","volume":"3","author":"Wu","year":"2008","journal-title":"PLoS One"},{"key":"2023062408113504800_btaa629-B45","doi-asserted-by":"crossref","first-page":"3113","DOI":"10.1016\/j.jmb.2017.08.013","article-title":"OPUS-DOSP: a distance- and orientation-dependent all-atom potential derived from side-chain packing","volume":"429","author":"Xu","year":"2017","journal-title":"J. Mol. Biol"},{"key":"2023062408113504800_btaa629-B46","doi-asserted-by":"crossref","first-page":"286","DOI":"10.1002\/pro.3327","article-title":"OPUS-CSF: a C-atom-based scoring function for ranking protein structural models","volume":"27","author":"Xu","year":"2018","journal-title":"Protein Sci"},{"key":"2023062408113504800_btaa629-B47","doi-asserted-by":"crossref","first-page":"3970","DOI":"10.1021\/acs.jctc.0c00186","article-title":"OPUS-Fold: an open-source protein folding framework based on torsion-angle sampling","volume":"16","author":"Xu","year":"2020","journal-title":"J. Chem. Theory Comput"},{"key":"2023062408113504800_btaa629-B48","doi-asserted-by":"crossref","first-page":"1359","DOI":"10.1021\/acs.jctc.9b01054","article-title":"OPUS-refine: a fast sampling-based framework for refining protein backbone torsion angles and global conformation","volume":"16","author":"Xu","year":"2020","journal-title":"J. Chem. Theory Comput"},{"key":"2023062408113504800_btaa629-B49","doi-asserted-by":"crossref","first-page":"16856","DOI":"10.1073\/pnas.1821309116","article-title":"Distance-based protein folding powered by deep learning","volume":"116","author":"Xu","year":"2019","journal-title":"Proc. Natl. Acad. Sci. USA"},{"key":"2023062408113504800_btaa629-B50","doi-asserted-by":"crossref","first-page":"W174","DOI":"10.1093\/nar\/gkv342","article-title":"I-TASSER server: new development for protein structure and function predictions","volume":"43","author":"Yang","year":"2015","journal-title":"Nucleic Acids Res"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btaa629\/33862466\/btaa629.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/36\/20\/5021\/50692656\/btaa629.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/36\/20\/5021\/50692656\/btaa629.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,6,24]],"date-time":"2023-06-24T23:34:20Z","timestamp":1687649660000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/36\/20\/5021\/5873143"}},"subtitle":[],"editor":[{"given":"Arne","family":"Elofsson","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2020,7,17]]},"references-count":50,"journal-issue":{"issue":"20","published-print":{"date-parts":[[2020,12,22]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btaa629","relation":{},"ISSN":["1367-4803","1367-4811"],"issn-type":[{"value":"1367-4803","type":"print"},{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2020,10,15]]},"published":{"date-parts":[[2020,7,17]]}}}