Molecular docking calculation is a computational chemistry technique that has been widely used to predict the most suitable binding structure between enzyme and inhibitor. The obtained information is helpful for structural modification of an inhibitor to enhance its biological activity. In this research, effect of using five different atomic charges on the accuracy of AutoDock Vina program for molecular docking calculation was investigated. 167 receptor-inhibitor complex structures were taken from the PDBbind database. Atomic charges were calculated using PM7 , Gasteiger-Marsili, MMFF9 4 , QEq, and QTPIE methods. The docked structures predicted from the calculations were then compared with the corresponding X-ray structures to examine the binding mode. RMSD values were also calculated to evaluate the accuracy by considering success rate of the agreement between predicted and experimental data. The results show that success rates of using atomic charges calculated with PM7, Gasteiger-Marsili, MMFF94, QEq, and QTPIE methods are 82.0%, 82.0%, 78.4%, 81.4%, and 80.8%, respectively. This indicate that PM7 and Gasteiger-Marsili atomic charges give the most accurate results. However, inspecting the RMSD values of each structure demonstrates that Gasteiger-Marsili atomic charges give the lowest RMSD values and are suitable for inhibitors with any size. In addition, the calculation time for this method is less than that of the PM7 . Therefore, the Gasteiger-Marsili atomic charges are recommended for the molecular docking calculation using AutoDock Vina.
