Genotoxic potential of a given chemical entity is an important toxicology endpoint in drug discovery. In vivo mutagenicity, as assessed by the AMES bacterial reverse mutation assay, is a good predictor of genotoxicity. Compounds that test positive in the experimental AMES test are considered to have high genotoxic potential and are usually not developed further. A reliable computational model of the experimental AMES test would result in time and resource savings. We investigated a mechanism-based computational approach and its ability to predict the genotoxic potential of aniline and aromatic amines from structure. The model computes the stability of the nitrenium ion directly, and from the acetate conjugate of aniline using density functional theory (DFT). The HOMO energy of the aniline species is used to assess its oxidative potential. A consensus model was built based on these mechanism-based approaches to predict the result of the AMES test. The resulting model was used for genotoxicity prediction in ongoing drug discovery projects. This model showed very good success and had greater than 80% accuracy at discriminating between AMES positive and negative compounds, prospectively. Retrospective studies of a historical dataset of 400 internal compounds were used to further refine the mechanism-based consensus prediction model. The model and analyses of its performance will be presented.

Mechanism-based modeling for predicting the result of the AMES test for aniline genotoxicity from chemical structure

248th National Meeting (2014)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.