期刊
TOXICOLOGY IN VITRO
卷 29, 期 7, 页码 1392-1404出版社
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.tiv.2015.05.017
关键词
Antimitotic agents; R-group QSAR; 2D-QSAR; Support vector machine; Kernel-based PLS; CoMSIA
类别
资金
- All India Council for Technical Education (AICTE), New Delhi
- Council of Scientific and Industrial Research (CSIR), New Delhi
- University Grants Commission (UGC), New Delhi
Antimitotic agents are potential compounds for the treatment of breast cancer. Cytotoxicity is one of the parameters required for anticancer activity. A validated comparative molecular modeling study was performed on a set of phenylindole derivatives through R-group QSAR (RQSAR), regression-based and linear discriminant analysis (LDA)-based 2D QSAR studies and kernel-based partial least square (KPLS) analyses as well as CoMSIA 3D-QSAR study. Antiproliferative activities against two breast cancer cell lines (MDA-MB-231 and MCF7) were separately used as dependent variables. The RQSAR analysis highlighted different E-state indices and pharmacophoric requirements of important substitutions. The best 2D-QSAR model is established on the basis of three machine learning tools - MLR, SVM and ANN. The 2D-QSAR models depicted importance of different structural, physicochemical and topological descriptors. While RQSAR analyses demonstrated the fingerprint requirements of various substitutions, the KPLS analyses showed these requirements for the entire molecule. The CoMSIA model further refines these interpretations and reveals how subtle variations in these structures may influence biological activities. Observations of different modeling techniques complied with each other. The current QSAR study may be used to design potential antimitotic agents. It also demonstrates the utilities of different molecular modeling tools to elucidate the SAR. (C) 2015 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据