Artificial intelligence in drug research and development: a review of methods and applications in drug repurposing

Abstract Artificial Intelligence (AI) plays an increasingly significant role in drug research and development, particularly in drug repurposing, which involves identifying new therapeutic indications for existing pharmacological compounds. From classical algorithms-based tools e.g. DrugRep and DrugRepo, to more recent innovations (e.g. RosettaVS, RepurposeDrugs), this paper reviews the latest AI methods (e.g. AlphaFold3, mediKanren, AdaDR, TxGNN) and their design applications in drug repurposing, particularly in machine learning, deep learning, and biological network analysis. It discusses strategies using predictive models, natural language processing, and big data analysis to accelerate the identification of promising candidates for clinical repurposing. Data availability, model interpretability, and results validation challenges are also highlighted. This review suggests that AI represents a groundbreaking tool in drug repurposing, ranging from protein structure prediction to knowledge graph reasoning, which can significantly reduce the time and costs of developing new therapies.