Prediction of Micronucleus Assay Outcome using in Vivo Activity Data and Molecular Structure Features

Abstract In vivo micronucleus assay is the widely used genotoxic test to determine the extent of chromosomal aberrations caused by the chemical compounds in human beings, which plays a significant role in the drug discovery paradigm. To reduce the uncertainties of the in vivo experiments and the expenses, we intended to develop novel machine learning-based tools to predict the toxicity of the compounds with high precision. A total of 472 compounds with known toxicity information were retrieved from the PubChem Bioassay database and literature. The fingerprints and descriptors of the compounds were generated using PaDEL and ChemSAR for the analysis. The performance of the models was assessed using three tires of evaluation strategies such as 5-fold, 10-fold, and external validation. The accuracy of the models during external validation lay between 0.57 and 0.86. Note that a combination of fingerprints and random forest showed reliable predictive capability. In essence, structural alerts causing genotoxicity of the compounds were identified using the structural activity relationship model of SARpy tool. This study highlights that the structural alerts such as chlorocyclohexane and trimethylamine are likely to be the leading cause of toxicity in humans, further validated using the Toxtree application. Indeed, the results from our study will assist in scrutinizing the genotoxicity of the compounds with high precision by replacing extensive sacrifice of animal models.