1Assistant Professor, CSE Department, Brainware University, Barasat, Kolkata – 125, West Bengal, India. Email: rbkpccst@gmail.com
2I.T.S Institute of Health & Allied Sciences Ghaziabad. Email: thangarajmpt@gmail.com
3JSS College of Pharmacy, Mysuru-570015. Email: rajath.1301@gmail.com
4Department Of Pharmaceutical Sciences, Rabindranath Tagore University, Hojai, Assam. Email: subhamghosh8927@gmail.com
5Research scholar at Assam University, Silchar. Email: djisu212@gmail.com
6Department of Electronics and Communication Engineering, School of Engineering and Technology, CGC University, Mohali – 140307, Mohali, Punjab, India. Email: ramandeep.j3335@cgcuniversity.in
7Associate Professor Community Medicine Department Malabar Medical College Hospital and Research centre, Calicut, Kerala Community Medicine. Email: speaktodrkrishnaraj@gmail.com
Artificial intelligence (AI) is reshaping drug discovery by improving how teams prioritize biological hypotheses, triage chemical space, and optimize leads under multi-parameter constraints. Across the early pipeline—from target identification and validation to hit discovery and lead optimization—modern machine learning integrates human genetics, multi-omics, biomedical knowledge graphs, structure prediction, and molecular representation learning to support decision-making in iterative design–make–test–analyze (DMTA) cycles. At the same time, the field faces persistent barriers that determine real-world impact: biased and noisy labels, inflated retrospective benchmarks due to leakage and overly permissive data splits, limited prospective validation, and insufficient uncertainty calibration. This review synthesizes methods and best practices across the target-to-lead workflow, focusing on (i) data foundations and representations, (ii) AI for target discovery and evidence integration, (iii) structure-enabled screening and interaction modeling, and (iv) multi-objective, synthesis-aware lead optimization using predictive and generative models. We conclude with practical recommendations for building credible, auditable AI workflows aligned with emerging regulatory guidance and good AI practice principles.
Keywords: Artificial intelligence; machine learning; target identification; virtual screening; docking; generative models; lead optimization; retrosynthesis; ADMET; DMTA
How to cite this article: Banerjee R, Thangaraj M, Raju RN, Ghosh S, Das J, Kaur R, Krishna Raj JS. Artificial Intelligence in Drug Discovery: Transforming Target Identification to Lead Optimization. Int J Drug Deliv Technol. 2026;16(16s): 65-74. DOI: 10.25258/ijddt.16.16s.8
Source of support: Nil.
Conflict of interest: None