Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{202359,
        author = {Vaidehi Iyengar},
        title = {AI-Driven Molecular Docking in Drug Repurposing: Methodologies, Tools, Case Studies, and Translational Challenges},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {7043-7050},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=202359},
        abstract = {Background and Objective: Drug development is among the most expensive and time-consuming processes in science, costing approximately $2.6 billion and 10–15 years per approved drug, with over 90% of candidates failing before approval. Drug repurposing — identifying new therapeutic uses for existing compounds — offers a faster, lower-cost alternative by leveraging established safety profiles. This review aims to critically evaluate the role of artificial intelligence (AI)-driven molecular docking in accelerating drug repurposing, examining current methodologies, landmark case studies, key databases, and translational challenges.
Methods: A comprehensive narrative review of peer-reviewed literature was conducted covering five principal AI methodologies: structure-based virtual screening (SBVS) with ML-enhanced scoring functions, ligand-based virtual screening (LBVS) and QSAR modelling, graph neural network (GNN)-based drug-target interaction prediction, AlphaFold 2 and 3-integrated structural proteomics, and knowledge graph/NLP-based network pharmacology. Landmark case studies (baricitinib for COVID-19; halicin as a novel antibiotic) and major databases (DrugBank, ChEMBL, PDB, AlphaFold DB) were systematically reviewed.
Results: AI-driven docking pipelines demonstrate clear superiority over traditional physics-based scoring functions in accuracy and throughput, enabling screening of over 100 million compounds in under 48 hours. AlphaFold 2 and 3 have expanded the druggable proteome to over 200 million predicted structures. Baricitinib, identified via BenevolentAI’s knowledge graph, reduced COVID-19 mortality by 38% in randomised trials and achieved WHO standard-of-care status. Halicin, discovered through deep learning, demonstrated novel antibacterial activity against drug-resistant pathogens in vivo. Key limitations identified include data bias, model interpretability (the black-box problem), the AlphaFold docking gap, benchmark reproducibility issues, and a persistent clinical translation gap.
Conclusion: AI-driven molecular docking has transformed drug repurposing from a serendipitous process into a systematic, high-throughput computational strategy with demonstrated clinical impact. Addressing current limitations through explainable AI, federated learning, benchmark standardisation, and integrated computational-experimental workflows will be critical to accelerating the translation of AI-predicted candidates into approved therapies.},
        keywords = {Drug Repurposing; Molecular Docking; Artificial Intelligence; Virtual Screening; AlphaFold; Graph Neural Networks},
        month = {May},
        }