QSAR based on multiple linear regression and PLS methods for the anti-HIV activity of a large group of HEPT derivatives

Quantitative structure-activity relationships have been developed for a set of 107 inhibitors of the HIV-1 reverse transcriptase, derivatives of a recently reported HIV-1 specific read: 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT). The activity of these compounds was investigated by means of multiple linear regression (MLR) and PLS regression techniques and topological indexes as well as several tabulated physicochemical substituent constants were used as predictor variables. The results obtained indicate that the anti-HIV activity of the HEPT derivatives is strongly dependent on hydrophobic factors as expressed by the Hansch constant (Sigma pi(R1+R2)), and especially dependent on the geometric factors mainly accounted for by the (1) chi(N)(R2) and (4) chi(p)(N) molecular connectivity indexes and also for the molecular volume (Vx), the Taft steric constant (Es(2R1)), and the Verloop parameter for the smallest width value (B1(3R1)). Besides, for this data set, comparison of the quality of MLR and PLS models show that PLS is a better approach to MLR for improving the interpretability of the data and also to exhibit models with a better predictive quality.