Nature Reviews Drug Discovery | Target Identification and Assessment in the Era of AI
InSilico MedicinePeer-Reviewed Publication
Target identification is a critical and challenging step in drug discovery, with only a small fraction of human genes considered druggable and even fewer successfully targeted by approved therapies. Traditionally, this process has been slow and complex, but artificial intelligence (AI) is transforming it into a more systematic, data-driven approach.
A recent review by Insilico Medicine highlights how AI is accelerating target discovery by integrating vast multimodal datasets—including omics data, clinical records, imaging, and scientific literature—to uncover novel disease mechanisms and therapeutic targets.
Advanced machine learning methods, such as supervised and unsupervised learning, graph neural networks, and generative AI models, enable researchers to prioritize targets, simulate biological systems, and generate new hypotheses with greater precision. Platforms like PandaOmics and emerging “virtual biologist” AI agents further enhance this capability by synthesizing complex biological knowledge.
Clinical progress demonstrates the real-world impact of these approaches, with AI-enabled platforms accelerating timelines and contributing to promising therapies, such as the TNIK inhibitor rentosertib for idiopathic pulmonary fibrosis.
Looking ahead, the field is moving toward AI-driven closed-loop systems that integrate computational predictions with automated experimentation, aiming to significantly improve efficiency, success rates, and the delivery of new treatments to patients.
- Journal
- Nature Reviews Drug Discovery