The efficacy of conceptual DFT descriptors and docking scores on the QSAR models of HIV protease inhibitors.

This study critically examines the role of conceptual DFT descriptors and docking scores on a diverse set of 156 inhibitors of HIV proteases. Five QSAR models were developed on the basis of available experimental IC(50) values (HIV-I and HIV-IIIB infected MT4 and CEMSS cells and HIV-I infected C8166 cells) and sixth QSAR model was generated by combining the inhibitors of all five models. B3LYP/6-31G(d) optimizations were carried out on all considered inhibitors, and the results are compared with more economic semi-empirical SCF AM1 results in order to find out the best and efficient way of descriptor calculations. Interestingly semi-empirical results appear to be satisfactory for this class of inhibitors. Selected QSAR models were validated by taking about 20% of inhibitors in the test sets. The effect of the number of descriptors on the R(2) and R(2)(cv) values was tested and three to four orthogonal descriptors based models were selected to be the optimum ones to avoid over correlation.