QSAR studies of bioactivities of 1-(azacyclyl)-3-arylsulfonyl-1H-pyrrolo[2, 3-b]pyridines as 5-HT₆ receptor ligands using physicochemical descriptors and MLR and ANN-modeling

Four molecular descriptors were selected from a pool of variables using genetic algorithm, and then used to built a QSAR model for a series of 1-(azacyclyl)-3-arylsulfonyl-1H-pyrrolo[2,3-b]pyridines as 5-HT₆ receptor agonists or antagonists, useful for the treatment of central nervous system disorders. Simple multiple linear regression (MLR) and a nonlinear method, artificial neural network (ANN), were used to model the bioactivities of the compounds; while MLR gave an acceptable model for predictions, the ANN-based model improved significantly the predictive ability, being more reliable for the prediction and design of novel 5-HT₆ receptor ligands. Topology and molecular/group sizes are important requirements to take into account during the development of novel analogs. Few physicochemical descriptors were used to model the bioactivities of a series of 1-(azacyclyl)-3-arylsulfonyl-1H-pyrrolo[2,3-b]pyridines as 5-HT₆ receptor inhibitors, giving predictive QSAR models, especially when using ANN as a nonlinear regression.