Modeling of cytochrome P-450 enzyme inhibitors activities using 2D/3D QSAR

Abstract

In this research, quantitative structure–activity relationship (QSAR) studies were carried out on the inhibitory activities of a set of nicotine derivatives against the cytochrome-p450 2A6 (CYP2A6) enzyme. Two-dimensional quantitative structure–activity relationship (2D-QSAR) models were developed using multiple linear regression (MLR) and linear square-support vector machine (LS-SVM) methods. The result of statistical parameters of the MLR method show that the correlation coefficient (R2) and standard error (SE) for the training set respectively are R2 = 0.702, SE = 0.49 and for the test set R2 = 0.689, SE = 0.52 and the results of statistical parameters of the LS-SVM method for the training set R2 = 0.993, SE = 0.10 and for the test set R2 = 0.977, SE = 0.20. The obtained results reveal the superiority of LS-SVM over MLR model. Then three-dimensional quantitative structure–activity relationship (3D-QSAR) model was developed using comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) on the same dataset of nicotine derivatives. The acquired statistical parameters of the CoMFA model for the training set are R2 = 0.884, SE = 0.316 and for the test set R2 = 0.847, SE = 0.33, F = 41.27, Q2 = 0.581, while the statistical values of CoMSIA model for the training set are R2 = 0.889, SE = 0.31 and for the test set R2 = 0.670, SE = 0.54, F = 39.049, Q2 = 0.554. The results of this study revealed that the CoMFA model was more predictive and could be helpful in designing novel potent nicotine derivatives with enhanced inhibitory activity.