QSAR study of the 5-HT1A receptor affinities of arylpiperazines using a genetic algorithm-artificial neural network model

Genetic algorithm-multiparameter linear regression (GA-MLR) and genetic algorithm-artificial neural network (GA-ANN) models have been used for prediction of the 5-HT1A receptor affinities (pK(i)) of 66 arylpiperazines. A large number of theoretical descriptors were calculated for each compound. The genetic algorithm (GA) was used for selection of the variables that resulted in the best fit to the MLR and ANN models. The models were generated using seven descriptors as variables. For evaluation of the predictive power of the models, pK(i) values of 13 compounds in the prediction set were calculated. Mean percentage deviation (MPD) for the GA-MLR and GA-ANN models were 0.344 and 0.065, respectively. Comparison of the results obtained by use of the models reveals the GA-ANN model is superior to the GA-MLR model.