Claude designs protein minibinders against scorpion neurotoxins
Researchers used Claude, an internal research model, to design minibinders targeting five scorpion neurotoxins. In collaboration with Citrini Research, they prompted Claude and followed up four times, resulting in 318,484 generated designs, 180 selected for testing, and an additional 31 designs, for a total of 211. Laboratory partners Adaptyv Bio and Twist Bioscience evaluated all designs and 17…
Key points
- Claude generated 318,484 protein designs, selected 180 for testing, and with the team added 31 more, totaling 211 designs.
- Adaptyv Bio and Twist Bioscience lab‑tested the designs and found 17 that met binder criteria, each against a single toxin.
- The campaign ran about 64 hours, used roughly 10,000 GPU hours, and produced two strongest binders (Cn2‑B1, Cn2‑B2) with single‑digit nanomolar KD.
The design campaign ran for about 64 hours and consumed roughly 10,000 GPU hours. Claude autonomously chose toxin structures from the Protein Data Bank and AlphaFold databases, identified hotspots, and employed 12 design tools. The two strongest binders, Cn2‑B1 and Cn2‑B2, showed single‑digit nanomolar dissociation constants and appear to compete with a known neutralizing antibody (LR). The team has not yet confirmed toxin neutralization in cellular or animal models, and they note that turning binders into safe, affordable antivenoms remains a major challenge. External experts Susana Vázquez Torres and Timothy P. Jenkins described the work as a promising first step.
The authors are releasing all 211 designs and associated data to encourage further research on antivenom development for neglected tropical diseases.
Claude designed protein binders against scorpion neurotoxins
anthropic.com · 20 October 2026
Loading the full article…
This text was published by anthropic.com and written by Anthropic. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
1source- Hacker News discussion · 41 pointsnews.ycombinator.com
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Research
All →- OpenAI brings back Aleksander Madry to focus on AI safety · 1 src
- Anthropic says small AI may self‑replicate by year‑end after experiment with GLM‑5.2 · 2 src
- ALTK-Evolve tops τ²-bench retail leaderboard using GPT-4.1 with learned guidelines · 2 src
- Stanford PhD uses generative AI to draft genetic blueprints for microscopic viruses · 1 src
- OpenAI claims AI-generated proof of quasi‑Riemann hypothesis · 1 src
Comments
via GitHub Discussions