Drug Dev Res. 2026 Apr;87(2):
e70270
Drug repurposing involves identifying new therapeutic applications for existing clinically evaluated compounds. In contrast to conventional drug development, which typically spans over a decade and demands substantial financial investment, repurposed drugs can achieve regulatory approval in approximately half the time and cost by capitalizing on their established pharmacokinetic, safety, and clinical profiles. This review provides a comprehensive analysis of the traditional and computational strategies employed in drug repurposing. Experimental methodologies include binding affinity assays, clinical data mining and phenotype-based screening. Computational approaches are categorized into structure-based, signature-based, pathway-based, knowledge-based, and target-based strategies. The recent integration of artificial intelligence (AI) and machine learning (ML) within repurposing pipelines is also examined, emphasizing their ability to efficiently process large-scale datasets, improve the predictive accuracy of drug-target interactions, and support the advancement of repurposing efforts. Furthermore, this review systematically compares prominent computational platforms, virtual screening tools, and bioinformatics resources, highlighting their respective strengths and limitations. Emerging AI-driven models, such as deep learning architectures, graph neural networks, knowledge graphs, and network pharmacology frameworks, have transformative roles in broadening the scope of drug repurposing. This comprehensive review is intended to assist medicinal chemists, computational biologists, and drug discovery scientists in expediting research efforts by effectively utilizing existing resources for repurposing-driven innovations.
Keywords: artificial intelligence; computational drug discovery; drug discovery; drug repurposing; translational research