The Generation and Validation of QSAR Model

QSAR modeling was performed on MOE. Initially 192 2D-descriptors were calculated for each compound of training set. The best descriptors were selected by using MOE “QuaSAR-Contingency” applications. Descriptors with contingency coefficient above 0.6 and Cramer’s, uncertainty, and correlation coefficients above 0.2 were selected for QSAR which proposed nine 2D-descriptors including apol (Sum of the atomic polarizabilities), bpol (Sum of the absolute value of the difference between atomic polarizabilities of all bonded atoms in the molecule), a_acc (Number of hydrogen bond acceptor atoms), a_heavy (Number of heavy atoms), logP(o/w) (Log of the octanol/water partition coefficient), logS (Log of the aqueous solubility (mol/L)), TPSA (topological polar surface area (Å2) c), Weight (molecular weight) and wienerPol (Wiener polarity number) as “best” for our dataset. These descriptors are widely used to predict biological activity and ADMET properties of dataset [46]. QSAR model was built by selecting the inhibitory activities of compounds (as descriptor variable) and calculated descriptors as model fields. Regression analysis was performed on the training set, and r² and RMSE values of the fit were obtained. The generated QSAR model was validated by cross validation by leave-one-out method. QSAR fit was used to validate the model by cross validation method. The predictive activities and the residual values of each compound of training set was assessed. Residual value, Z-score and predicted values were calculated for validation and cross-validation of model. The results were interpreted by drawing correlation plot between predicted values (actual activity) on x-axis versus predicted IC₅₀ values on y-axis. The developed model was used to predict the biological activities of test set and test set 2 compounds in terms of correlation coefficient (r²) between the experimental and predicted activities of test set. The developed QSAR model was used for the prediction of biological activities of selected Hits.