UK launches £8m project to build largest drug-protein interaction dataset for AI-led drug discovery

The UK has launched an £8m initiative to create what is claimed will be the world’s largest dataset of drug-protein binding data, as part of wider efforts to speed up AI-driven medicine development.

The OpenBind consortium, based at the Diamond Light Source in Oxfordshire, aims to generate more than 500,000 experimentally validated protein-ligand structures and measurements over five years—around 20 times more than have been publicly collected in the past 50 years.

Protein-ligand interactions show how potential drugs attach to proteins in the body, a critical process for determining whether a treatment is likely to work against a specific disease.

The Department for Science, Innovation and Technology’s Sovereign AI Unit is providing up to £8m in initial funding for OpenBind, announced during London Tech Week.

The consortium will bring together scientists from global institutions to create a high-quality data foundation for training AI models that can better predict how new drugs interact with disease targets.

Project leaders say the initiative could reduce the time and cost of bringing new medicines to market by as much as £100bn by improving research and development efficiency.

OpenBind will also collaborate with pharmaceutical and AI companies including Isomorphic Labs, Astex Pharmaceuticals, Genentech, and Chai Discovery, which are expected to take part in upcoming investment discussions to scale the initiative.

Sir Demis Hassabis, chief executive of Isomorphic Labs, said: “High-quality biochemical data supports superior AI models, which in turn helps us design new drug candidates faster.
“We’re delighted to partner with the OpenBind Consortium and the UK government to cultivate this vital resource. This is a brilliant initiative for UK science, and we’re proud to support it from its inception.”

In addition to pharmaceutical development, the dataset is expected to have applications in molecular and engineering biology, including the design of enzymes to help address environmental problems such as plastic waste.

Separately, Imperial College London has announced a partnership with the World Economic Forum to establish a Centre for AI-driven Innovation in the UK. The centre will join the Forum’s global network focused on the Fourth Industrial Revolution and will aim to support cross-sector collaboration on AI adoption.

Taken together, the announcements mark a wider push by the UK government to strengthen the country’s position in bioscience and artificial intelligence, using public-private partnerships to drive scientific innovation, attract investment and talent, and promote international collaboration.

Diamond Light Source, the UK’s national synchrotron science facility, uses intense beams of light to examine the structure of molecules at atomic resolution—critical to understanding how drugs interact with proteins at a molecular level.

The OpenBind initiative aims to address what its backers describe as a longstanding gap in pharmaceutical research, where limited high-quality data on drug-protein interactions has slowed progress in developing reliable AI tools for drug discovery.