Zuckerberg's project unveils AI model for drug development – Reuters

Biohub, a research organization founded by Meta CEO Mark Zuckerberg and his wife Priscilla Chan, has officially unveiled an AI-powered global model of protein biology. 

This was reported by Reuters. The new development is designed to radically accelerate and reduce the cost of creating highly effective drugs through digital modeling tools.

The model is built on the fourth generation of evolutionary large-scale modeling (ESM) and operates on the principle of large language models, but instead of human text, it analyzes biological protein sequences. The algorithm identified patterns in the structure of 6.8 billion proteins, allowing it to study the underlying rules of biology and predict the behavior of new compounds that do not yet exist in nature. Researchers have already tested the open-source system in practice, designing new protein interactions for cancer and immune targets that successfully reactivated immune cells to fight tumors during laboratory tests.

To accelerate global progress, the developers are providing free access to the technology on the biohub.ai platform, where non-commercial researchers will be allocated special computing credits. The product will also be integrated into leading global analytical platforms, including AWS Bio Discovery and SandboxAQ. This AI platform resulted from the merger of the Chan-Zuckerberg Initiative’s biomedical and technology divisions under the Biohub umbrella and the acquisition of the startup EvolutionaryScale; since 2015, the couple has allocated over $7 billion to scientific and charitable projects.

In 2020, for the first time in the history of medicine, a drug was created using artificial intelligence. And it took just one year to complete this work—five times faster than what doctors typically spend on developing new drugs.

Earlier, it was reported that Chinese artificial intelligence helped catch a murder suspect.