Mark Rackham, VP Chemistry at BenevolentAI will be giving a talk at the “Hatfield MedChem" meeting
An AI Drug Discovery Case Study: Discovery and Design of Neuroprotective Compounds for Treating ALS
Artificial intelligence can enable meaningful inferences to be made from huge bodies of data and therefore has the potential to augment many aspects of drug discovery, from target discovery through to patient stratification. The promise of AI has been lauded for many years, but examples of using it in end-to-end drug discovery remain scarce.
Herein we present a case study in the use of AI for drug discovery, where we used machine learning methods across the drug discovery pipeline. Building from our bespoke biomedical knowledge graph, we began by capturing the target identification strategy and logic of our discovery scientists as an algorithm, and systematically queried our knowledge graph for desirable polypharmacology. This workflow led to the discovery of novel neuroprotective molecules in a patient-derived model of ALS.
Starting from these tool molecules, we subsequently used predictive and generative machine learning models to drive compound design. This has resulted in a series of novel, highly efficient and CNS-penetrant molecules with efficacy in patient derived ALS assays. Our lead molecule was first synthesised 3 months from project inception, and the series is on track for candidate selection in Q3 2019.
Finally, we have trained predictive algorithms on complex clinical datasets and used the resulting models to stratify ALS patients, with the aim of running more robust clinical trials in appropriate patient groups. Although early in this workflow, we have demonstrated the power of machine learning to balance trials arms and drive more efficient trial design.
Mark is VP, Chemistry and Neurodegeneration Area Lead at BenevolentAI. Mark is responsible for driving the early discovery portfolio in the Neurodegeneration area, and for directing the development of AI tools across our chemistry programs. Before joining BenevolentAI, Mark worked for Pharma (GSK) and Biotech (Heptares) both in the UK and US, working on novel chemical modalities as therapies in Immuno-Oncology, Inflammation and Neurodegeneration.