Anthropic's Claude AI has independently designed protein binders targeting deadly scorpion neurotoxins, which kill over 3,250 people annually. Working largely autonomously for 24 hours inside the Claude Science research environment, the model generated 318,484 designs against five toxins from four scorpion species after an initial prompt. The research team subsequently steered the work with additional instructions, leading to 211 tested designs where 17 met the criteria for a binder.
Laboratory testing revealed that the two strongest candidates bound extremely tightly to Cn2, a toxin from the Mexican scorpion. These binders appear to dock near the site of an antibody known to neutralize Cn2 in mice, though their effectiveness as an antivenom remains untested. External experts Susana Vázquez Torres and Timothy P. Jenkins praised the project as a promising first step, with independent results confirmed by two separate labs.
This milestone builds upon an earlier study where Claude designed mini-binders against 14 of 15 target proteins, achieving hit rates up to 35 percent. Such computational biology breakthroughs highlight the expanding role of advanced artificial intelligence in medical research and protein engineering.