Proteins designed using artificial intelligence (AI) can block the lethal effects of toxins found in cobra, adder, and other deadly snake venoms.
AI-designed proteins could form the basis of a new generation of treatments for snakebites. An estimated 100,000 people die from snakebites each year, and they still receive much the same treatment today as they did 100 years ago.
The study, published January 15 in Nature, also demonstrates how machine learning has enhanced the field of computational protein design. Tasks that previously took months, years, or even proved impossible, such as designing a new protein that successfully blocks another protein, can now be accomplished in seconds.
Scientists are using AI to devise innovative new proteins
“It’s scary,” says Joseph Jardine, an immunologist at Scripps Research in La Jolla, California. “We went from being like, ‘We can’t do this,’ to working on a proof of concept to solve a real problem.”
Snake bites are a significant cause of death and permanent disability in many parts of the world. The World Health Organization in Geneva, Switzerland, has listed snakebites as a top neglected tropical disease, along with dengue fever and rabies.
But treatments have remained largely unchanged for more than a century, mostly based on antibodies in serum taken from horses and sheep immunized with snake venom. These antivenoms vary in safety and effectiveness, and their usefulness is limited because they must be administered in clinics by trained staff, says the Toxin Research Institute at the University of Costa Rica, Clodomiro Picad in San José. Scholar José María Gutierrez points out.
triple the amount of toxins
Developing a snakebite treatment wasn’t on David Baker’s radar, a computational biophysicist at the University of Washington in Seattle, when his lab announced an innovative protein design program called RFdiffusion in late 2022. Ta. Like DALL-E and Midjourney, this program has proven adept at designing small proteins that bind tightly to target proteins, including those associated with cancer and autoimmune diseases.
Susana Vázquez Torres, a biochemist in Baker’s lab, was interested in tackling neglected diseases and wondered if RF diffusion could help improve snakebite treatment. Snake venom consists of various protein toxins that cause paralysis and tissue damage.
AI dreamed up a blizzard of new proteins. Does any of it actually work?
Vázquez-Torres, Baker and colleagues designed a “mini-binder” that uses RF diffusion to recognize key regions of three types of toxins produced by snakes in the serpent family, including cobras, mambas, and spotted snakes.