Construction of pioneering quantitative structure activity relationship screening models for abuse potential of designer drugs using index of ideality of correlation in monte carlo optimization.

Drug abuse is a worldwide wide problem affecting individual, society and the environment in general and it is nothing less than the attempted ecocide. Designer drugs are the chemical substances used for recreational purposes and have addictive properties. The production of designer drugs at disturbing pace is creating difficulties for the investigators in their testing. Computational evaluation method can be an interesting approach for early checking of abusive drugs. In the present work, quantitative structure activity relationship (QSAR) models are developed for abusive potential of designer drugs using SMILES and graph based parameters. Dopamine transporter/serotonin transporter inhibition (DAT/SERT) ratio was used as endpoint and the whole data set was divided into eight non identical splits for development of the models using balance of correlation technique of Monte Carlo optimization. The internal and external cross validation results confirmed that the models created with index of ideality of correlation were reliable and robust in prediction. The developed models followed all the five principles of the Organisation for Economic Co-operation and Development. The best model split 2 possessed good fitting ability and internal as well as external predictive ability and it was used in explanation of activity trends of different classes of designer drugs.