The document discusses applying computational models to problems in toxicology, drug discovery, and beyond. It summarizes recent work using machine learning models and other in silico techniques to predict drug-induced liver injury (DILI) and interactions with transporters like hOCTN2. Models were able to classify compounds as DILI-positive or negative with over 75% accuracy when tested on external datasets. The techniques discussed could help prioritize compounds for further testing and filter libraries to avoid reactive or toxic features.