QSAR Prediction of Octanol-Water Partition Coefficients for Potential Dental Monomers

Objectives: The purpose of this study was to develop a model for the prediction of logP_o/w values for potential dental materials. By predicting the logP_o/w value, the toxicity of potential materials under study can be evaluated and the most promising candidate selected. Methods: The AM1 semiempirical quantum mechanical method is used to perform conformational searches for a training set of molecules for which experimental data is available. CODESSA was used to compute descriptors from the quantum mechanical data. A QSAR (quantitative structure property relationship) was derived a model that related the logP_o/w to chemically meaningful descriptors via a multilinear equation and statistical analysis was performed on the resulting equation. Results: Some 92 molecules of various functionalities were included in the training used to derive the model. Three descriptors were obtained: molecular surface area, total dipole moment of the molecule, and fractional atom charge weighted partial positive surface area), all of which can be rationalized as related to the effect under study. The result of the statistical analysis is: R² = 0.945, adjusted R² = 0.943, R_cv² = 0.940; the variance inflation factor (VIF) for the three descriptors above are 1.116, 1.044, and 1.162 respectively. Conclusions: This QSAR model can be used to predict the logP_o/w value for a wide variety of organic functional groups that may be used as dental materials.