AIToday
Large Language ModelsAI in HealthcareTHE DECODERPublished: Oct 21, 2026, 19:01 JST

Claude designs 318,484 scorpion antivenom candidates

Send on LINE
Claude designs 318,484 scorpion antivenom candidates

3 Key Points

  1. What happened

    Claude generated 318,484 protein binder designs targeting five toxins from four scorpion species, working on its own for 24 hours after an initial prompt. Of 211 designs tested in the lab, 17 met the criteria for a binder.

  2. Why it matters

    The lab results suggest that AI can independently produce viable protein binder candidates, a step toward faster early-stage drug discovery, though whether they work as an antivenom has not been tested.

WHO IT HITSDrug-discovery researchers and biotech teams gain a proof point that an AI environment like Claude Science can autonomously generate lab-testable protein designs, potentially shortening early candidate screening.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

The scorpion work follows an earlier study in which Claude designed mini-binders against 14 of 15 target proteins, achieving hit rates of 22 to 35 percent compared to the typical 10 to 15 percent. That prior result provides context for the current effort: the same AI environment, Claude Science, produced a large batch of designs and a smaller set of lab-validated binders. The research team steered the work with additional instructions after an initial 24-hour autonomous run, suggesting a human-in-the-loop process rather than full automation. External experts Susana Vázquez Torres and Timothy P. Jenkins described the work as a promising first step, and two independent labs confirmed the results, which may lend credibility to the findings. However, whether the binders actually function as an antivenom has not been tested, leaving the practical medical payoff uncertain.

FAQ
How many designs did Claude test in the lab?
Of 318,484 generated designs, Claude picked 180 candidates and the team added 31 more, for 211 designs tested in the lab.
What did external experts say about the work?
Susana Vázquez Torres and Timothy P. Jenkins called it a promising first step, and two independent labs confirmed the results.
Has Claude designed protein binders before?
In an earlier study, Claude designed mini-binders against 14 of 15 target proteins, with hit rates of 22 to 35 percent compared to the typical 10 to 15 percent.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleRicoh Japan adds monthly GPU server rental to on-prem LLM kit