Single-Sequence Deep Learning Delivers Crystal-Quality Models of Covalent K-Ras G12 Hotspot Complexes - PubMed
2 hours ago
- #KRAS
- #Deep Learning
- #Covalent Inhibitors
- Chai-1 predicts covalent K-Ras(G12C) complex structures accurately without MSA.
- Achieves pocket-aligned RMSDs < 2 Å for diverse inhibitors (ARS-853 to BBO-8520).
- Successfully models binding poses for K-Ras(G12D) and K-Ras(G12S) inhibitors.
- Limitations include handling leaving groups, bond properties, and stereochemistry.
- Provides ~40-fold higher throughput than AlphaFold3 with comparable accuracy.
- Offers an efficient tool for accelerating covalent drug discovery beyond cysteine.