Objective

The Design-Make-Test-Analyse cycle (DMTA cycle) is central to the process of constructing and testing hypotheses in drug discovery. It is used to iteratively improve the overall profile of therapeutically relevant compounds. For an average project hundreds of DMTA cycles are required to improve the weakly active compounds from high throughput screening (HTS) to the very compound that meets the criteria for a candidate drug. Even today, almost all steps in the DMTA cycle are done manually with many people and various compound hand-overs between the disciplines involved. Very often this leads to cycle times of several weeks and thus to 6-8 years from project start to begin of clinical trials. AstraZeneca’s ambition is to fully automate the DMTA cycle and thereby to reduce the time between compound design and receipt of test data from currently 3-4 weeks to less than one day. Novel compounds will be automatically designed, synthesised, purified and analysed against integrated bio-assays, using Machine Learning to decide autonomous optimisation cycles of chemical structural modifications to drive the discovery of medicines of the future.