HelixDock AI-Assisted Virtual Drug Screening Platform Launches: Candidate Compound Screening 60 Times Faster

HelixDock AI-Assisted Virtual Drug Screening Platform Launches: Candidate Compound Screening 60 Times Faster

HelixDock, a virtual-screening platform jointly developed by UK AI drug-discovery company Exscientia and the University of Oxford's Department of Chemistry, has officially launched. The platform combines 3D protein-structure prediction (AlphaFold 3) with molecular-dynamics simulation, compressing candidate-compound virtual screening from a conventional 6 months to 3 days.

Traditional drug screening requires synthesizing thousands of compounds and running wet-lab tests; the full cycle takes 6 to 18 months and is expensive. HelixDock uses deep-learning models to predict binding affinity between compounds and targets, filtering out 99 percent of invalid candidates before synthesis.

Initial projects target two sites: tau protein in Alzheimer's disease and alpha-synuclein in Parkinson's disease. Fergus McKenzie, who leads the Oxford team, said traditional screening for these two hard-to-drug targets would have taken years; with HelixDock, the team locked in four high-potential candidate compounds within three months.

Exscientia plans to open the platform to small and mid-sized biotech firms worldwide on a pay-per-use basis. The company expects HelixDock to serve more than 80 research institutions in 2026.