The production of new drugs is complex and time-consuming, but the use of artificial intelligence (AI) has the potential to reduce costs and time. Maryam Astero, a doctoral researcher at the Centre for Young Synbio Scientists, is focusing on this area.
In computational biology, after identifying a target molecules (e.g. new drug, enzym) with a desired property, the challenge is to determine how to generate this molecule at a reasonable cost using commercially available starting materials. This question can be answered by retrosynthesis, which is the deconstruction of the target molecule into simple or commercially available starting materials. However, retrosynthesis is costly, time-consuming, and multi-step process requiring expert chemists with deep knowledge of chemical reactions.
Computer-aided synthetic planning and retrosynthesis prediction have gained considerable attention in recent years. The advancement of graph neural networks (GNN) has led to new methods for retrosynthesis. In this approach, atoms are represented by nodes, and there is a link between two nodes if there is a covalent bond between the corresponding atoms.
A collaborative effort among machine learning specialists, biochemists, and biologists has the potential to significantly accelerate the retrosynthetic analysis of new target molecules, ultimately leading to faster delivery of new drugs to the market. By harnessing deep learning strategies in the drug discovery process, it becomes possible to reduce the time and cost involved in developing new treatments and therapies, ultimately benefiting human health.
The development of new drugs and therapies through more efficient and cost-effective methods contributes to UN Sustainable Development Goals 3: Good Health and Well-being, Goal 9: Industry, Innovation and Infrastructure, and Goal 17: Partnerships for the Goals.