SO3LR Tauri: ML force field trained on 148M structures brings quantum accuracy to biomolecules across the periodic table
DaveJuergens · x · 2026-11-06
A new arXiv paper (2611.06237) presents SO3LR Tauri, a family of machine-learned force fields aiming to describe all major classes of (bio)molecules—proteins, DNA, lipids, carbohydrates—at quantum-mechanical accuracy.
- Couples a fast SO3krates network with explicit physics modules: pair-specific short-range repulsion, environment-dependent dispersion, and electrostatics
- Trained on 148 million structures spanning 85 elements from several quantum-chemical datasets
- Accurately reproduces nucleobase-pair and charged protein-ligand interactions, carbohydrate conformer ordering, and molecular dipoles against coupled-cluster references; resolves transition pathways of organic and enzymatic reactions
- Dynamical simulations recover experimental solvent densities, ion hydration structures, amino-acid pKa values, and the structure/dynamics of proteins, membranes, nucleic acids, glycoproteins, and metalloproteins without system-specific fitting
- Released in three sizes, reaching 0.7–3.2 μs/day simulation speeds
More from Research
- 30 years on, Lean turns Edinburgh's 'proofs beside code' idea into reality — Sanity · 2026-11-06
- RAX From Harvard/NVIDIA/Berkeley Lifts Robot Success From 26% to 94% With Frozen Weights — jiqizhixin · 2026-11-06
- Stanford Releases ScholarCatalyst, a Retrieval Benchmark Built From Researchers' Judgments — StanfordAILab · 2026-11-06
- Open-Weights Robot-Use Harness: Qwen3.8 Scores 31.39, Beating GPT-6 Astra's 28.97 — aigclink · 2026-11-06
- Autonomous AI R&D Runs 'Extremely Dumb Experiments', Researcher Reports — panickssery · 2026-11-06
- ICLR reviews starting to leak, authors still waiting on scores — Invariant_n_Cauchy · 2026-11-06