In this paper, an attempt was made to develop a Quantitative Structure Activity Relationship (QSAR) model on a series of quinazoline derivatives acting as Protein tyrosine kinases (erbB-2) inhibitors using Multiple Linear Regression, Principal Component Regression and Partial Least Squares Regression methods. Among these three methods, Multiple Linear Regression (MLR) method has come out with a very promising result as compared to other two methods. Various 2D descriptors were calculated and used in the present analysis. For model validation, the dataset was divided into training and test sets using spherical exclusion method. The developed MLR- QSAR model was found to be statistically significant with respect to training (r2 =0.956), cross-validation (q2 = 0.915), and external validation (pred_r2= 0.6170). The developed MLR model suggests that Estate Contribution descriptors SaaOE-Index (30.07%) and SsCIE-index (15.79%) are the most important descriptors in predicting Tyrosine kinase (erbB-2) inhibitory activity. Electron withdrawing group at 4th position of quinazoline enhances the activity as evident by positive value of SsClE-index (15.79). In addition, for quinazoline substituents, estate contribution descriptors SsCH3E – index has a large deactivating effect.
Keywords: Protein tyrosine kinase (erbB-2), multiple linear regression (MLR), principal component regression (PCR), partial least squares (PLS), quinazoline, Partial Least Squares Regression, cross-validation, external validation, SaaOE-Index, SsCIE-index
Rights & PermissionsPrintExport