Show simple item record

dc.contributor.authorHwang, Howook
dc.contributor.authorVreven, Thom
dc.contributor.authorWhitfield, Troy W.
dc.contributor.authorWiehe, Kevin
dc.contributor.authorWeng, Zhiping
dc.date2022-08-11T08:07:59.000
dc.date.accessioned2022-08-23T15:38:11Z
dc.date.available2022-08-23T15:38:11Z
dc.date.issued2011-08-01
dc.date.submitted2013-02-22
dc.identifier.citation<p>Proteins. 2011 Aug;79(8):2467-74. doi: 10.1002/prot.23070. Epub 2011 Jun 1. <a href="http://dx.doi.org/10.1002/prot.23070">Link to article on publisher's site</a></p>
dc.identifier.issn0887-3585 (Linking)
dc.identifier.doi10.1002/prot.23070
dc.identifier.pmid21633973
dc.identifier.urihttp://hdl.handle.net/20.500.14038/25883
dc.description.abstractProteins often undergo conformational changes when binding to each other. A major fraction of backbone conformational changes involves motion on the protein surface, particularly in loops. Accounting for the motion of protein surface loops represents a challenge for protein-protein docking algorithms. A first step in addressing this challenge is to distinguish protein surface loops that are likely to undergo backbone conformational changes upon protein-protein binding (mobile loops) from those that are not (stationary loops). In this study, we developed a machine learning strategy based on support vector machines (SVMs). Our SVM uses three features of loop residues in the unbound protein structures-Ramachandran angles, crystallographic B-factors, and relative accessible surface area-to distinguish mobile loops from stationary ones. This method yields an average prediction accuracy of 75.3% compared with a random prediction accuracy of 50%, and an average of 0.79 area under the receiver operating characteristic (ROC) curve using cross-validation. Testing the method on an independent dataset, we obtained a prediction accuracy of 70.5%. Finally, we applied the method to 11 complexes that involve members from the Ras superfamily and achieved prediction accuracy of 92.8% for the Ras superfamily proteins and 74.4% for their binding partners.
dc.language.isoen_US
dc.relation<p><a href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=pubmed&cmd=Retrieve&list_uids=21633973&dopt=Abstract">Link to Article in PubMed</a></p>
dc.relation.urlhttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC3341935/
dc.subject*Artificial Intelligence
dc.subjectProtein Structure, Secondary
dc.subjectProteins
dc.subjectAmino Acids, Peptides, and Proteins
dc.subjectBioinformatics
dc.subjectComputational Biology
dc.subjectSystems Biology
dc.titleA machine learning approach for the prediction of protein surface loop flexibility
dc.typeJournal Article
dc.source.journaltitleProteins
dc.source.volume79
dc.source.issue8
dc.identifier.legacycoverpagehttps://escholarship.umassmed.edu/bioinformatics_pubs/25
dc.identifier.contextkey3761406
html.description.abstract<p>Proteins often undergo conformational changes when binding to each other. A major fraction of backbone conformational changes involves motion on the protein surface, particularly in loops. Accounting for the motion of protein surface loops represents a challenge for protein-protein docking algorithms. A first step in addressing this challenge is to distinguish protein surface loops that are likely to undergo backbone conformational changes upon protein-protein binding (mobile loops) from those that are not (stationary loops). In this study, we developed a machine learning strategy based on support vector machines (SVMs). Our SVM uses three features of loop residues in the unbound protein structures-Ramachandran angles, crystallographic B-factors, and relative accessible surface area-to distinguish mobile loops from stationary ones. This method yields an average prediction accuracy of 75.3% compared with a random prediction accuracy of 50%, and an average of 0.79 area under the receiver operating characteristic (ROC) curve using cross-validation. Testing the method on an independent dataset, we obtained a prediction accuracy of 70.5%. Finally, we applied the method to 11 complexes that involve members from the Ras superfamily and achieved prediction accuracy of 92.8% for the Ras superfamily proteins and 74.4% for their binding partners.</p>
dc.identifier.submissionpathbioinformatics_pubs/25
dc.contributor.departmentProgram in Bioinformatics and Integrative Biology
dc.contributor.departmentDepartment of Biochemistry and Molecular Pharmacology
dc.source.pages2467-74


This item appears in the following Collection(s)

Show simple item record