A QSAR classification study on inhibitory activities of 2-arylbenzoxazoles against cholesteryl ester transfer protein

作者：Chen Meimei; Lai Xinmei; Yang Xuemei^*

来源：Medicinal Chemistry Research, 2014, 23(4): 1878-1886.

DOI：10.1007/s00044-013-0789-8

• QSAR
• Classification
• 2-Arylbenzoxazoles
• SVM
• LDA

Inhibition of cholesteryl ester transfer protein (CETP) is considered as a new therapy for treating coronary heart disease. Recently, a series of compounds with a 2-arylbenzoxazole skeleton have been identified as potent CETP inhibitors, however, some of them can also be noninhibitors. Therefore, it is highly desirable to have models that can predict whether a compound interacts with CETP. In this work, a quantitative structure-activity relationship study was developed for modeling and classifying CETP inhibitory activity for a dataset of 84 2-arylbenzoxazoles using diverse structural descriptors. The stepwise linear discriminant analysis (LDA) method was then used to explore the best descriptors responsible for inhibitory activity. The LDA and nonlinear support vector machine (SVM) method were employed to establish classification models. These models were strictly validated using internal and external approaches. The SVM model gave the prediction accuracy of 92.65 % on the training set and 81.25 % on the test set, as well as that of the LDA model 82.35 and 75 %. In addition, petitjeanSC, PEOE_VSA_NEG, E_nb, and vsurf_EWmin1 were found to have high correlation with the CETP inhibitory activity. Overall, the SVM model can act as a fast filter for designing new 2-arylbenzoxazoles with better CETP inhibitory activity in future.

• 出版日期2014-4
• 单位福建中医药大学