BioEmu2 simulates peptide folding & binding in 40 GPU-minutes vs 10+ GPU-years for MD — 100,000x speedup
FrankNoeBerlin · x · 2026-10-20
Frank Noe's team demonstrates BioEmu2 simulating the dynamics of an intrinsically disordered peptide folding and binding to its protein receptor via an intermediate state.
The same 5-millisecond simulation would take over 10 GPU-years with conventional MD simulation; BioEmu2 needs just 40 minutes on a single GPU at 5-microsecond steps — a >100,000x speedup.
More from Research
- AI Trained on 40M Face Photos Predicts Cancer Survival From a Single Selfie, per Lancet Digital Health — EricTopol · 2026-10-20
- CyberArena benchmarks AI agents in live attack-and-defense combat — 2077AIDataFoundation · 2026-10-20
- PNAS study: brief AI chats drive 17,950 people to sign petitions and donate, attitudes don't predict behavior — _FelixSimon_ · 2026-10-20
- Citrini Used Anthropic's Internal Model to Design Scorpion Antivenom Proteins Overnight — citrini · 2026-10-20
- The Explainability Illusion: ViT patch heatmaps are a training artifact, not model property — ziv_ravid · 2026-10-20
- Agentic RL data scaling laws paper: environments beat tasks and trajectories, 1.7x compute savings — omarsar0 · 2026-10-20