{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/accurate-de-novo-prediction-of-protein","title":"Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model","arxiv_id":"1609.00680","date":"2016-09-02","proceeding":null,"authors":["Sheng Wang","Siqi Sun","Zhen Li","Renyu Zhang","Jinbo Xu"],"abstract":"Recently exciting progress has been made on protein contact prediction, but\nthe predicted contacts for proteins without many sequence homologs is still of\nlow quality and not very useful for de novo structure prediction. This paper\npresents a new deep learning method that predicts contacts by integrating both\nevolutionary coupling (EC) and sequence conservation information through an\nultra-deep neural network formed by two deep residual networks. This deep\nneural network allows us to model very complex sequence-contact relationship as\nwell as long-range inter-contact correlation. Our method greatly outperforms\nexisting contact prediction methods and leads to much more accurate\ncontact-assisted protein folding. Tested on three datasets of 579 proteins, the\naverage top L long-range prediction accuracy obtained our method, the\nrepresentative EC method CCMpred and the CASP11 winner MetaPSICOV is 0.47, 0.21\nand 0.30, respectively; the average top L/10 long-range accuracy of our method,\nCCMpred and MetaPSICOV is 0.77, 0.47 and 0.59, respectively. Ab initio folding\nusing our predicted contacts as restraints can yield correct folds (i.e.,\nTMscore>0.6) for 203 test proteins, while that using MetaPSICOV- and\nCCMpred-predicted contacts can do so for only 79 and 62 proteins, respectively.\nFurther, our contact-assisted models have much better quality than\ntemplate-based models. Using our predicted contacts as restraints, we can (ab\ninitio) fold 208 of the 398 membrane proteins with TMscore>0.5. By contrast,\nwhen the training proteins of our method are used as templates, homology\nmodeling can only do so for 10 of them. One interesting finding is that even if\nwe do not train our prediction models with any membrane proteins, our method\nworks very well on membrane protein prediction. Finally, in recent blind CAMEO\nbenchmark our method successfully folded 5 test proteins with a novel fold.","url_abs":"http://arxiv.org/abs/1609.00680v6","url_pdf":"http://arxiv.org/pdf/1609.00680v6.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"accurate-de-novo-prediction-of-protein","repo_url":"https://github.com/j3xugit/RaptorX-Contact","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"protein-folding","task_name":"Protein Folding"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1609.00680","atlas_url":"https://app.syntology.ai/?focus=1609.00680","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}