Quantitative structure-activity relationship modeling of antioxidant activities of hydroxybenzalacetones using quantum chemical, physicochemical and spatial descriptors.

We have modeled antioxidant activities of hydroxybenzalacetones against lipid peroxidation induced by t-butyl hydroperoxide (pC1), gamma-irradiation (pC2) and also their 1,1-diphenyl-2-picryl hydrazyl (DPPH) free radical scavenging activity (pC3) using quantitative structure-activity relationship technique. The quantitative structure-activity relationship models were developed using different statistical methods like stepwise multiple linear regression, genetic function approximation and genetic partial least squares with descriptors of different categories (quantum chemical, physicochemical, spatial and substituent constants). The models were validated by internal validation and randomization techniques. The model predictivity was judged on the basis of their cross-validated squared correlation coefficient (Q2) and modified r2 (r m 2) values. The best models for the two responses, pC1 and pC2, were obtained by genetic partial least squares technique while the best model for the third response, pC3, was obtained by genetic function approximation technique. The developed models suggest that the distribution of charges on the phenolic nucleus and the phenolic oxygen as well as the charged surface areas of the molecules together with the geometry and orientation of the substituents significantly influence all the three types of responses (pC1, pC2 and pC3). The developed models may be used to design hydroxybenzalacetones with better antioxidant activities.