A quantitative structure-activity relationship (QSAR) study of dermal absorption using theoretical molecular descriptors.

Quantitative structure-activity relationship (QSAR) models were developed for the prediction of dermal absorption based on experimental log Kp data for a diverse set of 101 chemicals obtained from the literature. Molecular descriptors including topostructural (TS), topochemical (TC), shape or three-dimensional (3D) and quantum chemical (QC) indices were calculated. Based on this information, a generic predictive model was created using the diverse set of 101 compounds. In addition, two submodels were prepared for subsets of 79 cyclic and 22 acyclic chemicals. A modified Gram-Schmidt variable reduction algorithm for descriptor thinning was followed by regression analyses using ridge regression (RR), principal components regression (PCR) and partial least squares regression (PLS). The RR results were found to be superior to PLS and PCR regressions. The cross-validated correlation coefficients for the full set and subsets were 0.67-0.87. Computational methods such as QSAR modelling can be used to augment existing data to prioritise chemicals that need to be studied further for toxicological evaluation and risk assessment.