Google DeepMind helped find drugs for liver fibrosis among existing medications

Geneticist Gary Peltz at Stanford University School of Medicine used Google DeepMind's Co-Scientist system to accelerate the hunt for medicines that can slow, stop, or reverse liver fibrosis. Peltz asked Co-Scientist to propose three drug candidates, while he identified two himself based on literature prominence. All five were tested on live human liver cells. Peltz's two picks showed no benefit, but two of Co-Scientist's three candidates blocked fibrosis and promoted regeneration. One of these drugs had only been linked to fibrosis in a few papers. Co-Scientist's standout pick was the cancer drug vorinostat, which blocked 91% of a damage response driving liver scarring. Co-Scientist's suggestions pointed toward drugs that reshape gene activity rather than targeting single pathways. The research was published in Advanced Science.