The DFT-B3LYP method, with the basis set 6-31G (d, p), was employed to calculate some quantum chemical
descriptors of 33 biphenyl imidazole derivatives as bombesin receptor subtype-3 agonists. The descriptors were then employed
to establish a quantitative structure activity relationship (QSAR) using combination of principal component analysis
(PCA) and radial basis function neural network (RBF). The statistical results indicate that the correlation coefficient
(R2) and cross validation using leave-one-out were 0.968 and 0.963, respectively. To validate the predictive power of the
resulting model, external validation was carried out on the test set. The results show that the PCA-RBF model has not
only favorable estimation stability but also good prediction power. Furthermore, it can be concluded that the agonist activity
of studied compounds toward the bombesin receptor subtype-3 depends on the electronic distribution.
Keywords: Bombesin receptor subtype-3 agonists, Density functional theory, Radial basis function neural network, QSAR,
Biphenyl imidazole derivatives, Nonlinear relationship.
Rights & PermissionsPrintExport