2-(2-Arylidenehydrazinyl) thiazol-4(5H)-ones having epidermal growth factor receptor (EGFR) inhibitory activity were subjected to quantitative structure activity relationship (QSAR) (2 and 3 dimensional) analysis and pharmacophore study in the present work. The best 2 dimensional (2D) QSAR model had r2 (squared correlation coefficient), q2 (cross validated correlation coefficient) and pred_r2 (predictive correlation coefficient) values of 0.8630, 0.7652 and 0.9533 respectively with Partial least square regression (PLSR) analysis. This model showed that count of number of nitrogen atoms separated from oxygen atom by four bonds (T_N_O_4) and RadiusOfGyration (size descriptor) descriptors contribute positively and count of number of any atoms separated from carbon atom by four bonds (T_T_C_4) contributes inversely to the biological activity. The k-nearest neighbor (kNN) method produced a significant 3 dimensional (3D) QSAR model exhibiting q2 and pred_r2 values of 0.6615 and 0.8833 respectively. PLSR gave a significant 3D QSAR model having r2, q2 and pred_r2 values of 0.8449, 0.7816 and 0.7834 respectively. Both 3D QSAR models depict the need of less bulky substituents at R2 position while PLSR 3D QSAR model also portrays the need of more bulky substituent at R4 position. The identified common pharmacophore features are one aromatic (AroC) and three hydrogen bond acceptors (HAc) obtained from Molsign and Pharmagist approaches. The present work may be useful for further lead optimization and designing of potent anticancer agents.
Keywords: EGFR, QSAR, Pharmacophore, kNN, Pharmagist.