Advancements and Challenges of Artificial Intelligence in Drug Discovery and Synthetic Organic Chemistry

AI-based research has progressed rapidly with the widespread use of AI modelling tools and the increase in electronic data in chemistry. In the field of medicinal chemistry, AI has been used for several decades to improve the success rate of drug discovery, with a long history in in-silico screening, prediction of mechanism of action and physical properties such as solubility. The authors have developed AI models with the aim of building our own platform for AI-based drug discovery. In the field of synthetic organic chemistry, databases of chemical reactions collected from patents and literature are used to train AI models for proposing new reactions, predicting reaction yields, etc. In this review, I will introduce our AI models for drug discovery, which include molecular design, activity prediction and ligand binding pose prediction. As the latest trend of AI development in synthetic organic chemistry, recent studies on discovering new chemical reactions and predicting reaction yields are presented. In addition, the challenges of AI in drug discovery and synthetic organic chemistry are discussed for the further development of AI studies in chemistry.