Google DeepMind Co-Scientist for Liver Fibrosis Drug Repurposing

Google DeepMind's Co-Scientist AI has demonstrated the ability to identify repurposed medicines to treat liver fibrosis, outperforming human selection in a controlled laboratory testbed. This capability allows researchers to uncover hidden connections in vast biomedical literature to find drugs that can slow, stop, or reverse liver scarring.

AI-Driven Drug Discovery for Liver Fibrosis

Liver fibrosis is a scarring process that can progress to cirrhosis, contributing to more than 1.4 million deaths annually. To combat this, Professor Gary Peltz of the Stanford University School of Medicine utilized Co-Scientist to scan existing medical literature and propose candidate drugs for repurposing.

In a study published in Advanced Science, the research team compared the efficacy of drugs selected by the AI against those selected by a human expert:

• Human Selection: Professor Peltz identified two candidate drugs based on their frequent appearance in liver fibrosis literature. Neither drug showed benefit against fibrosis in the lab tests.
• AI Selection: Co-Scientist proposed three candidate drugs. Two of these successfully blocked fibrosis and promoted the regeneration of liver cells.

One of the AI-selected drugs had been linked to liver fibrosis in only a few scientific papers, illustrating the system's ability to find "needles in the haystack" of scientific literature.

Technical Impact of Vorinostat

Among the AI's recommendations, the cancer drug vorinostat emerged as a standout candidate. In experimental tests using live human liver cells, vorinostat blocked 91% of a damage response that typically drives liver scarring.

Shift Toward Gene Activity Modulation

The discovery process revealed a fundamental difference in how the AI identifies potential treatments compared to traditional methods. While many traditional approaches target a single fibrosis pathway, Co-Scientist suggested drugs that reshape overall gene activity.

Professor Peltz suggests that these types of drugs, which modulate gene activity rather than targeting a single pathway, represent a promising new generation of anti-fibrotic medicines.

Co-Scientist feels like a collaborator that’s read everything available about biomedical science, with the reasoning capabilities to find the connections that we’re currently missing. — Professor Gary Peltz, Stanford University School of Medicine