NVIDIA's BindClaw uses agents for protein design: 15 of 18 binders hit calmodulin
sokrypton · x · 2026-10-29
NVIDIA's Digital Biology team unveiled BindClaw, an agentic protein design system that starts from a prompt or paper, defines the biological mechanism, builds mechanism-specific computational assays, and iteratively designs binders. Agents evaluate designs against biological objectives like competing with native partners and preferring molecular states. In a human-guided calmodulin campaign, 15 of 18 designs bound calcium-loaded CaM—including subnanomolar binders validated by both BLI and SPR—while 13 of 15 showed no binding to calcium-depleted CaM. Code will be released in a few weeks.
More from coding & agent
- Anthropic API Key Leaked; Agents Detected and Rotated It Within 15 Minutes — tylerjdunn · 2026-10-30
- Coding agents seem to have converged on the same PR format, sparking weight-training jokes — mattpocockuk · 2026-10-30
- Vercel Plugin Arrives in Google Antigravity's New Plugin Marketplace — steren · 2026-10-30
- Shopify execs require working prototypes in meetings, built with Codex — OpenAIDevs · 2026-10-30
- Developer uses Codex to map neighborhoods and vet communities while house-hunting — OpenAIDevs · 2026-10-30
- Mitchell Hashimoto runs Codex agents overnight to compare tools and prototype — OpenAIDevs · 2026-10-30