AI agents promise to transform drug discovery, but where do they really add value? Dr Eric Ma examines search, model choice, vendor lock-in and why scientific expertise matters more than ever.
A drug can form exactly the complex it was designed to make and still fail to work. Two recently approved induced proximity drugs show why what happens after complex formation – not formation alone – can decide the outcome.
Agentic AI, ultralarge virtual screening, molecular glues and human-relevant models were among the key developments in early drug discovery during Q3 2026. Drug Target Review examines what changed, why it matters and what to watch heading into Q4.
AI agents can search, analyse and plan across drug discovery workflows, but where should scientists hand over control? We examine what to automate, what to check and where human judgement matters most.
Insilico Medicine has nominated ISM1354, an AI-designed small-molecule GIPR antagonist, as a preclinical candidate for obesity and related metabolic disorders.
A new industry consortium led by Ginkgo Datapoints and Apheris is assembling a 10,000-antibody dataset to train AI models that can flag developability problems earlier in the drug discovery pipeline.
A mathematical model developed by researchers from Mass General Brigham and the University of Cyprus suggests that the experimental immunomodulatory antibody-drug conjugate HE-S2 may become progressively more effective as it reduces tumour size.
A newly identified epitope on the circumsporozoite protein of Plasmodium vivax could offer a route to vaccines that prevent liver-stage infection and curb the relapse-driven transmission cycle.
Five leading cancer researchers from across genomics, proteomics, cell biology, chemical biology and gene therapy dig into what’s driving cancer drug discovery forward and what’s still holding progress back.
AI agents promise to transform drug discovery, but where do they really add value? Dr Eric Ma examines search, model choice, vendor lock-in and why scientific expertise matters more than ever.