Google DeepMind's London laboratory has announced a potential breakthrough in treating a rare genetic disorder, using its AlphaFold protein-structure prediction system to identify a drug candidate that has shown remarkable efficacy in early clinical trials.
The drug, designated DM-4891, targets a misfolded protein responsible for Fabry disease, a painful and often fatal condition affecting roughly 1 in 40,000 people. Phase I trials showed the compound correcting protein misfolding in 87 per cent of patients with minimal side effects.
"This is exactly the kind of application we envisioned when we first demonstrated AlphaFold," said Demis Hassabis, DeepMind's co-founder and CEO. "We're not just predicting protein structures — we're using that knowledge to design medicines that work."
The announcement sent ripples through the pharmaceutical industry, where traditional drug discovery processes typically take 10-15 years and cost over £2 billion per successful compound. DeepMind's approach reduced the initial identification phase from years to weeks.
British biotech stocks rallied sharply on the news, with the FTSE All-Share Pharmaceuticals index rising 3.2 per cent. Analysts at JPMorgan described the results as "potentially transformative" for the economics of rare disease drug development.
