Poster: Combining Structural Alerts and Machine Learning Algorithms to Predict Human Molecular Initiating Events

Molecular Initiating Events (MIEs) are important chemical-biological interactions that start Adverse Outcome Pathways (AOPs). MIEs provide good targets for computational modelling as they are right on the boundary between biology and chemistry, and any computational models produced for them do not skip over large amounts of biological complexity, compared to trying to predict responses at an organ or organism level. We have developed in silico models for the prediction of important human MIEs for use in risk assessment using chemical understanding. The first of these approaches uses structural alerts, chemical fragments of known binders, to predict MIEs. The second of these approaches involves the use of machine learning algorithms fed with chemical information to predict MIEs. Combining these orthogonal in silico approaches provides further confidence in their predictions and a larger impact for computational toxicology.

The poster was presented by Tim Allen (University of Cambridge).

View the poster here.


Comments are closed.