The Governing Council of the University of Toronto

Published purpose

AI-enabled drug discovery has led to the rapid discovery of novel drug candidates and sparked significant investment in specialized drug companies in Canada and around the world. However, one important limitation of these proprietary platforms is they require large training datasets that are labor intensive to produce, and computationally expensive to analyze. This project will develop a self-driving lab (SDL) to develop a machine-learning (ML) algorithm that integrates experimental constraints directly into generative models as well as an orchestrator tool to control experiments performed and refine hypotheses tested. This approach not only streamlines the discovery process but also enhances the diversity and quality of candidates generated. The project team anticipates accelerated discovery by significantly reducing the time required to identify promising peptide drug candidates through automated, intelligent experimental design. This will lead to innovative therapeutics, as AI-driven hypothesis generation to discover peptides with unprecedented therapeutic potential.