Development of quantitative structure-activity relationships (QSAR) for alkylated polycyclic aromatic hydrocarbons to predict physicochemical properties

Anthony Skach

Petros Dimitriou-Christidis

Dr. Robin Autenrieth

Department of Civil Engineering

Texas A&M University

        This study of polycyclic aromatic hydrocarbons (PAHs) developed quantitative structure-property relationships through computational analysis using Cerius² v.4.9 molecular simulation software (Accelrys Inc., San Diego, CA) and ChemOffice (CambridgeSoft Corp., Cambridge, MA). Four physicochemical properties including solubility, vapor pressure, the octanol-water partition coefficient and Henry’s Law constant were collected for thirty-four PAHs including their methylated structures. Vapor pressure and solubility were converted to their liquid state formats calculated from their fugacity ratios. These converted physicochemical properties with their corresponding PAH and drawn molecular structures were arranged in a structures database and entered into the Cerius² QSAR+ software where molecular descriptors, including those that are topological, spatial, and thermodynamic, were calculated. The Genetic Function Approximation (GFA) algorithm generated ninety-nine QSAR equations for each physicochemical property using intricate regression methods.  Before selecting the best possible equation for each property, correlations between descriptors within the generated equations were analyzed to determine if equations could be simplified. The final QSPR equations yielded the following linear R² : 0.927 for log P_v,L , 0.992 for log H_c , 0.979 for log K_ow , and 0.954 and 0.931 for log S_L in units of milligrams per liter and millimoles per cubic meter respectively. These models were tested to test their validity with literature physicochemical property values of well-characterized compounds and used to predict these properties of PAHs with little or no experimental measurements.

Power Point Presentation: