2015. 56, 5 UDC 541.6:548.737 QSAR STUDIES OF HEPT DERIVATIVES AS ANTI-HIV DRUGS USING THE RASMS METHOD J. Tong1,2, X. Zhao1,2, L. Zhong1,2, J. Chang1,2 1College of Chemistry and Chemical Engineering, Shaanxi University of Science and Technology, Xian, P. R. China E-mail: jianbotong@sust.edu.cn Received January, 6, 2014 Random sampling analysis on the molecular surface (RASMS) is used to describe the chemical structures of 35 HEPT derivatives as anti-HIV drugs. <...> Here a quantitative structure activity relationship (QSAR) model is built by multiple linear regression (MLR). <...> The estimation stability and prediction ability of the model are strictly analyzed by both internal and external validations. <...> The correlation coefficients of the established MLR model, leave-one-out (LOO) crossvalidation, and predicted values versus experimental ones of external samples were r2 = 0.851, 2 DOI: 10.15372/JSC20150506 Keywords: random sampling analysis on molecular surface (RASMS), HEPT derivatives, quantitative structure activity relationship (QSAR), multiple linear regression (MLR). <...> QSAR studies play a fundamental role in predicting the biological activity of new compounds and identifying ligand-receptor interactions [13—15 ]. <...> The first step in constructing the QSAR models is to find one or more molecular descriptors © Tong J., Zhao X., Zhong L., Chang J., 2015 QCV = 0.746, and r2(test) = 0.815 respectively. <...> Structural descriptors have been classified into different categories according to different approaches, including physiochemical, constitutional, geometrical, topological, and quantum chemical descriptors. <...> The RASMS method derived from the atomic probe of protein (APP), the pseudoreceptor accessible surface (PRAS), and the approach to aimed area by random sampling on the molecular surface (ARSMS) was used to express the drug structures and biological activities with a multiple linear regression (MLR) model of 35 HEPT derivatives as anti-HIV drugs. <...> The proposed method was evaluated by predicting the activities of the derivatives in this paper; the results indicated that it is a useful tool for the investigation of drug QSAR. <...> Since drug targets are often protein and active peptides, there are eight different types of hybrid atoms from the amino acid serving as probes. <...> To characterize these probe atoms <...>