Quantitative Structure Activity Relationship (QSAR) studies using CoMFA and CPSA parameters were carried out on tetranortriterpenoids exhibiting diverse antifeedant activity to obtain models with high predictive ability. Satisfactory results with good statistical parameters were obtained from both methods. CoMFA gave a reasonably better predictive model with a cross-validated r2 value of 0.654 and a conventional r2 value of 0.953 for all compounds; however the analysis with CPSA parameters yielded better results only when the compounds were analyzed as individual classes based on their geometry. Single equation involving the chosen descriptors, based on their r2 values, for each class of compounds was calculated and a correlation was drawn between the experimental and calculated values, corroborated by good statistical parameters. The enhancement in the antifeedant activity of the photo products and the reduction in the activity of the microwave modified compounds are well brought out from the above studies. The present analysis reveals that better prediction models using CPSA parameters can be obtained by class wise analysis of the compounds.
Keywords: QSAR, tetranortriterpenoids, CoMFA, neem, CPSA