Quantitative Structure Property Relationship (QSPR) analysis was applied to 36 Persistent Organic Pollutants (POPs) using a combination of 0D, 1D, 2D and 3D molecular descriptors obtained by Semi empirical (pm3) method. The computed descriptors were correlated with the log of their experimental octanol-air partition coefficient (pKOA).Genetic function approximation was used to derive the most statistically significant QSPR model as a calibration model for predicting the pKOA of this class of molecules. Among the obtained QSPR models, the most statistically significant one was a five parameter linear equation with the squared correlation coefficient R2 value of 0.9889, adjusted squared correlation coefficient R2adj value of 0.9860 and Leave one out (LOO) cross validation coefficient (Q2) value of 0.9827. An external set was used for confirming the predictive power of the model (R2pred. = 0.7471). It is envisaged that the QSPR results identified in this study will offer an efficient and cost effective method of assessing the fate of POPs in the environment.
