中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
COMSAT: Residue contact prediction of transmembrane proteins based on support vector machines and mixed integer linear programming

文献类型:期刊论文

作者Zhang, Huiling; Huang, Qingsheng; Bei, Zhendong; Wei, Yanjie; Floudas, Christodoulos A.
刊名PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS
出版日期2016
英文摘要In this article, we present COMSAT, a hybrid framework for residue contact prediction of transmembrane(TM) proteins, integrating a support vector machine (SVM) method and a mixed integer linear programming(MILP) method. COMSAT consists of two modules: COMSAT_SVM which is trained mainly on position-specific scoring matrix features, and COMSAT_MILP which is an ab initio method based on optimization models. Contacts predicted by the SVM model are ranked by SVM confidence scores, and a threshold is trained to improve the reliability of the predicted contacts. For TM proteins with no contacts above the threshold, COMSAT_MILP is used. The proposed hybrid contact prediction scheme was tested on two independent TM protein sets based on the contact definition of 14 angstrom between C alpha-C alpha atoms. First, using a rigorous leave-one-protein-out cross validation on the training set of 90 TM proteins, an accuracy of 66.8%, a coverage of 12.3%, a specificity of 99.3% and a Matthews' correlation coefficient (MCC) of 0.184 were obtained for residue pairs that are at least six amino acids apart. Second, when testedon a test set of 87 TM proteins, the proposed method showed a prediction accuracy of 64.5%, a coverage of5.3%, a specificity of 99.4% and a MCC of 0.106. COMSAT shows satisfactory results when compared with 12 other state-of-the-art predictors, and is more robust in terms of prediction accuracy as the length and complexity of TM protein increase.
收录类别SCI
原文出处http://onlinelibrary.wiley.com/doi/10.1002/prot.24979/full
语种英语
源URL[http://ir.siat.ac.cn:8080/handle/172644/10182]  
专题深圳先进技术研究院_数字所
作者单位PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS
推荐引用方式
GB/T 7714
Zhang, Huiling,Huang, Qingsheng,Bei, Zhendong,et al. COMSAT: Residue contact prediction of transmembrane proteins based on support vector machines and mixed integer linear programming[J]. PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS,2016.
APA Zhang, Huiling,Huang, Qingsheng,Bei, Zhendong,Wei, Yanjie,&Floudas, Christodoulos A..(2016).COMSAT: Residue contact prediction of transmembrane proteins based on support vector machines and mixed integer linear programming.PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS.
MLA Zhang, Huiling,et al."COMSAT: Residue contact prediction of transmembrane proteins based on support vector machines and mixed integer linear programming".PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS (2016).

入库方式: OAI收割

来源:深圳先进技术研究院

浏览0
下载0
收藏0
其他版本

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。