Accelerating discovery of liver disease mechanisms
DeepMind Blog · 2026-05-16
DeepMind's Co-Scientist AI helped University of Edinburgh researchers identify the NLRP3 inflammasome as the molecular mechanism explaining why the approved MASH drug resmetirom only benefits a narrow patient subset, a hypothesis that was subsequently verified experimentally.
Appears in
Extraction
Topics: co-scientistliver-diseasedrug-discoveryai-for-sciencemash
Claims
- Co-Scientist synthesized evidence across liver biology and pharmacology to identify combination therapy candidates for MASH, a disease where single-target drugs fall short.
- Co-Scientist generated the hypothesis that the NLRP3 inflammasome is the specific molecular bridge coupling inflammation and metabolism in MASH, explaining resmetirom's limited efficacy.
- The NLRP3 inflammasome hypothesis had never previously been assembled into a single actionable explanation before Co-Scientist produced it.
- The AI-generated hypothesis was later experimentally verified and could enable targeted dual-therapies for MASH.
Key quotes
Co‑Scientist tackled a live, practical question: Why does the drug resmetirom – a recently approved treatment prescribed for a specific stage of MASH – only help a narrow slice of those eligible patients?
The system produced a hypothesis pinpointing the NLRP3 inflammasome as the specific molecular bridge coupling inflammation and metabolism in the disease — a connection never previously pulled together into a single, actionable explanation.