AI-designed proteins tackle century-old problem — making snake antivenoms

Proteins designed using artificial intelligence (AI) can block the deadly effects of toxins in the venom of cobras, vipers and other deadly snakes.

AI-designed proteins could form the basis of a new generation of therapies for snakebites, which kill about 100,000 people each year and are still treated as they were a century ago.

The study, published in Nature January 15¹is also a demonstration of how machine learning has energized the field of computational protein design. Challenges that previously took months, years, or even impossible—like designing a new protein to block another—can now be accomplished in seconds.

“It’s scary,” says Joseph Jardine, an immunologist at Scripps Research in La Jolla, California. “We moved from ‘we couldn’t even do that’ to proof-of-concept work solving real-world problems. »

In many parts of the world, snakebites are a significant cause of death and permanent disability. The World Health Organization in Geneva, Switzerland, has designated snakebite a high priority neglected tropical disease, alongside other illnesses such as dengue and rabies.

However, treatments have changed little in more than a century: most are based on antibodies present in the blood serum of horses and sheep immunized with snake venom. These antivenoms vary in safety and effectiveness and must be administered in a health center by trained personnel, which limits their usefulness, notes José María Gutiérrez, a toxinologist at the Clodomiro Picado Institute at the University of Costa Rica in San Jose.

Triple dose of toxin

Developing treatments for snakebites wasn’t on the radar of David Baker, a computational biophysicist at the University of Washington in Seattle, when his lab unveiled a revolutionary protein design program called RFdiffusion in late 2022. Like DALL-E and Midjourney, the program has proven adept at designing small proteins that bind strongly to target proteins, including those linked to cancer and autoimmune diseases.

Susana Vázquez Torres, a biochemist in the Baker lab, was interested in combating neglected diseases and wondered whether RF diffusion could help improve treatments for snakebites. Snake venoms are composed of various protein toxins that cause paralysis and tissue damage.

Vázquez Torres, Baker and their colleagues used RFdiffusion to design “mini-binders” that recognize key regions of three types of toxins produced by elapid snakes – the family that includes cobras, mambas and vipers.