In Silico Toxicology

Using state of the art tools inlcuding NVIDIA Titan V powered GPU based deep learning models we are building new models to predict toxicological outcomes. A key question for us in “how can a computer understand what a molecules looks like?” We are exploring new methods for molecular representation and new machine learning paradigms to build predictive relationships between those representations and toxicological effects. Chemical mutatgenicity is a key area of interest for chemical regulators world-wide. Out laboratory collaborates with our Federal regulator of chemicals to expidite the assessment and classification of chemicals in the Australian context. This work has world-wide ramifications for chemical regualtion practice.