Ke-Xian Chen, Hai-Ying Xie and Zu-Guang Li*

College of Chemical Engineering and Materials Science, Zhejiang University of Technology, 18 Chaowang Road, Hangzhou 310014, PR China

* Corresponding author: E-mail: lzg@zjut.edu.cn
Phone/Fax: +86-571-88320306

Abstract
Quantitative structure-activity relationship studies were carried out on some novel HCV NS5B polymerase inhibitors comprising 1,1-dioxoisothiazoles and benzo [b] thiophene-1,1-dioxides using genetic function algorithm (GFA) and molecular field analysis (MFA) techniques. The statistically significant 2D/3D-QSAR models (r² > 0.975) showed the indispensable structural requirements to improve the activity of this class. High r² CV values of 0.961 and 0.945 and r² pred values of 0.856 and 0.992 respectively for 2D/3D-QSAR models indicated the significant predictive ability of derived models. The validation of the models was done by full cross validation tests and external test set prediction. The results obtained can be exploited for modifications of the anti-HCV NS5B polymerase activity of this class of analogs.

Keywords: Hepatitis C virus NS5B polymerase inhibitors, 1,1-dioxoisothiazole, benzo[b]thiophene-1,1-dioxide, quantitative structure-activity relationship (QSAR), genetic function algorithm (GFA), molecular field analysis (MFA)