Quantitative structure-activity relationship of some pesticides.

Herbicides (benzodiazepinediones), insecticides (dioxatricyclododecenes) and larvicides (N-oxalyl derivatives of tebufenozide) have been quantitatively investigated to explore the relationship between the molecular structure and their biological activity using molecular operating environment (MOE) software. The study provides good predictive models, cross-validated by leave-out-one method (Loo). The positive contribution of the descriptor n-O (count of oxygen atom) suggests the additional oxygen atom substitution at R1 position, in addition to benzodiazepine moiety is favorable for herbicidal activity, whereas the negative contribution of y component of dipole moment (Dip(y)) indicates that electronic interactions are also crucial for the activity. The negative correlation of V(SA)2 and globularity (Glo) descriptors clearly indicates that the volume, shape, and rigidity of tebufenozide derivatives determine their larvicidal activity. The biparametric model for insecticides shows that the indicator variable l(CH-CH3) and R(PC) (negative partial charge) are detrimental for its activity. Most of the active compounds in the series have shown less value for these descriptors. The derived QSAR models also provide valuable insights to optimize their toxicity, which remains a major concern for environment safety.