Predictive models of Cannabinoid-1 receptor antagonists derived from diverse classes

Chemical database design is an important consideration for screening processes in drug discovery. More specifically, classification of a diverse compound set deeply influences the validation and the predictive power of prediction model for the designing of novel compounds. In this work, we investigated the effect of the reasonable classification on the prediction model. We first collected the known Cannabinoid-1 receptor antagonists. Following this, we calculate the chemical descriptors in order to classify the collected compounds. Finally, we build two predictive models via the 3D-QSAR using different molecular alignment and the alignment independent Molecular Interaction Field models.