Generating reliable synthetic routes to complex targets is a major challenge in drug discovery. A key part of this is identifying suitable reaction conditions cost-efficiently and in a reasonable timeframe. One synthetic transformation of particular interest to fragment-based drug discovery are C-H functionalisation reactions due to their pronounced regioselectivity and wide applicability.

Sam will be utilising machine learning techniques to generate robust, transferable models for C-H functionalisation that generalise across multiple targets, transformations, and datasets. This project will leverage in house data to construct and validate models fit for integration into Astex’ state-of-the-art synthesis platform.