Polycyclic aromatic hydrocarbons are toxic, carcinogenic and are widely distributed in the environment. Accurate prediction of their aqueous Henry’s law constant will be of immense help to environmental scientists in determining the fate of these chemicals in the environment. In this study, a Genetic function approximation (GFA)-QSPR analysis of some selected poly aromatic hydrocarbons (PAHs) was performed using different molecular descriptors. Five models for predicting the HLC of PAHs were generated. A seven parameter model with R2 = 0.996, R2adj = 0.994, Q2 = 0.989,R2 - Q2 = 0.007, R2pred. = 0.758, r2 – r02 / r2 =0.00, r2 – r‘ 02 / r2 = 0.00, K = 0.998, K’ = 1 was selected as the optimization model based on statistical significance. The Euclidean based applicability domain for training and test set compounds hinted that all the compounds fell within the applicability domain of the optimum QSPR model. The molecular descriptors; BCUTp-1l, ETA_Beta_ns, nFRing, topoDiameter, DPSA-2, LOBMIN, WD.mass were found to have profound influence on the predictive ability of the model. It is envisioned that the model will found excellent application in the prediction of Henry’s law constant of poly aromatic hydrocarbons that fall within its applicability domain.
