Polychlorinated aromatic compounds represent a large group of industrial and byproduct compounds which are resistant to chemical and biological degradation and highly toxic. QSAR analysis was performed on 74 molecules of three classes of polychlorinated aromatic compounds (polychlorinated dibenzo-p-dioxin (PCDDs), polychlorinated dibenzofuran (PCDF) and polychlorinated biphenyl (PCB)). A large number of about 1700 molecular descriptors was obtained from DFT (B3LYP/6-311+G*) level of calculation for each molecule and used in Genetic function algorithm (GFA) approach to generate 5 models, out of which the one with the highest statistical significance (Model-1: R2 = 0.9673, R2adjusted = 0.9592, R2cv = 0.9402, R2pred. = 0.7209, F-test = 118.48, LOF = 0.4377) was selected as the best. From the model generated, it seems to be very clear that polarizability, SP-7, ETA_Epsilon_5, GRAVH_3, and MOMI-R contribute positively to the toxicity of these compounds while MaxHBint5, ETA_dApha_B, ETA_Epsinlon-2, n5Ring and GRAV_2 contribute negatively. This validated model brings important insight to aid the prediction and identification of other toxic polychlorinated aromatic compounds.
