QSAR STUDY ANTICANCER DRUGS OF LETROZOLE AND VOROZOLE

Quantitativestructure-activity relationship (QSAR) studies were performed on twenty seven Letrozole and Vorozole derivatives, using multiple linear regression (MLR), , genetic algorithm (GA) and artificial neural networks (ANN), which were then used to construct the QSAR approches. A high predictive ability was observed for the MLR-MLR, MLR-ANN, and GA-ANN methods, with root mean sum square errors (RMSE) of 0.219, 0.0331, and 0.0216 respectively (N=20). The results obtained by using the GA-ANN model indicated that the activity of derivatives of Letrozole and Vorozole compounds was dependent on different parameters, including R5m, BIC0, nDB, EEig14r, Mor14e, Mor26v, EEig12r descriptors. In summary, a comparison of the quality of the ANN with different MLR methods demonstrated that ANN possesseda better predictive ability.