Uncovering repurposed medicines to fight liver fibrosis
DeepMind Blog · 2026-05-16
Stanford geneticist Gary Peltz published findings in Advanced Science showing that DeepMind's Co-Scientist outperformed expert human drug selection for liver fibrosis, with two of three AI-proposed candidates—including the cancer drug vorinostat, which blocked 91% of a key fibrotic damage response—proving effective in lab tests.
Appears in
Extraction
Topics: co-scientistliver-fibrosisdrug-repurposingai-for-scienceepigenetics
Claims
- Co-Scientist identified two effective drug repurposing candidates for liver fibrosis out of three proposed, while the researcher's own two expert-selected candidates showed no benefit.
- The cancer drug vorinostat, Co-Scientist's top recommendation, blocked 91% of a key damage response that drives liver scarring in lab tests with live human liver cells.
- One of Co-Scientist's successful picks had only appeared in a handful of liver fibrosis papers prior to this study, representing a needle-in-a-haystack discovery.
- Co-Scientist tended to select drugs that reshape gene activity broadly rather than targeting a single fibrosis pathway, a strategy Peltz argues deserves serious clinical consideration.
- The research was published in the journal Advanced Science.
Key quotes
His two drug picks showed no benefit against fibrosis. By contrast, of the three drug candidates selected by Co-Scientist, two blocked fibrosis and promoted the regeneration of liver cells.
One of these drugs had only been linked to liver fibrosis in a handful of papers – a needle in the haystack of scientific literature.
In Peltz's experiments, it blocked 91% of a damage response that can drive liver scarring.