Stanford's ProteoCraft hits 82.9% experimental hit rate on PD-L1 protein design
Lianhuiq · x · 2026-11-04
Stanford researchers unveiled ProteoCraft, a tree-search-based approach to hallucination-driven protein design. Experimental results: 29/35 designed PD-L1 miniproteins bound successfully (82.9% hit rate, 100% expression success), with a best measured SPR KD of 0.879 nM (best fitted 1.15 nM). Weaker results on IFNAR2: 16.7% hit rate for miniproteins (best KD 359 nM) and 40% for VHHs (best KD 157 nM). The strong low-nanomolar binding on PD-L1 highlights the promise of combining tree search with generative models for computational protein design.
Related event: Stanford Releases ProteoCraft, a Tree-Search Protein Design Framework(2 posts)→
More from Research
- TLAPS-Bench: 265 real-world TLA+ proof targets; top model GPT-6.1 Sol solves only 44.3% — tianyin_xu · 2026-11-04
- StreamDuo pairs a commentator with a wiki research agent to commentate unseen games — Zixian Gao · 2026-11-04
- Redwood researcher: AI research automation and runaway recursive self-improvement can come apart — deanwball · 2026-11-04
- UMAP 0.6 adds Adam and momentum layout optimizers, becoming defaults in 0.7 — leland_mcinnes · 2026-11-04
- UMAP 0.6 preview release candidate lands on PyPI with major changes — leland_mcinnes · 2026-11-04
- Training microscopic deformations in a sheet of glass to classify images using nothing but light — louisvarge · 2026-11-04