Modelling Vasorelaxant Activity of Some Drugs/Drug Candidates Using Artificial Neural Networks

Cardiovascular diseases are the most common health problems in developed and developing societies and the vasodilating agents are one of the medicinal groups to improve the life style of the patients suffering from the cardiovascular diseases. To study the quantitative structure-activity relationship of a number of pharmacological agents, the published data sets containing more than 10 vasodilating agents assessed on rat thoracic aorta, were collected from the literature. Different physico-chemical and structural descriptors of the compounds were computed using HyperChem registered (12 descriptors) and Dragon software (1479 descriptors). The more suitable descriptors (Jhetv, Lop, SP20, RDF020u, RDF030m and R6m) were selected using a combination of linear regression and genetic algorithm methods. The artificial neural networks method was used for modelling-log of vasodilating activity (pECSO) using selected descriptors. The statistical analyses were performed using SPSS software and the average percentage deviation between calculated and observed values for predicted data points studied in this work was 15.0 ( plus or minus 18.8).