Development of Computational Approaches for the Prediction of Regioselectivity and the Likely Products of Xenobiotic Metabolism

Abstract

Metabolism plays a vital role in the safety and efficacy of drugs and other xenobiotics. Computational approaches can be used to predict how xenobiotics will be metabolized in the human body, thereby guiding, for example, the drug development process. This dissertation describes the development and evaluation of computational methods to predict the metabolically labile atom positions of molecules as well as the chemical structures of the metabolites. The use of molecular alignment as the basis for predicting cytochrome P450 regioselectivity was investigated and a new approach was developed. Further, new methods were developed to generate the chemical structures of metabolites and to effectively prioritize these generated metabolites.