Pretrained protein models fail to learn 15N chemical shifts, study shows
KevinKaichuang · x · 2026-11-20
Kevin Kaichuang highlights that pretrained protein models still struggle with biophysics, notably failing to learn 15N chemical shifts. The accompanying work builds rBMRB: from 15,981 BMRB protein entries, condition filtering (pH, temperature) and re-referencing via LACS/PANAV statistics yield 6,123 clean entries, used to train the SLAE-CS ensemble model for 15N chemical shift prediction with applications to ligand binding, mutational effects, catalysis and dynamics.
Related event: Study Finds Protein Pretrained Models Fail to Learn 15N Chemical Shifts(2 posts)→
More from Research
- Interactive lessons record a 1.5B model's attention and token probabilities live — techartist_ · 2026-11-20
- Zero-cost MLM-head rescaling stabilizes SPLADE training, yet GTE-MLM remains unbeaten — tomaarsen · 2026-11-20
- After evaluating Forge for MTG, researcher writes his own engine for RL agents — petrusenko_max · 2026-11-20
- Study: AI agents barely learn social norms and rely mainly on being punished themselves — ghadfield · 2026-11-20
- Xeophon discusses the ECI benchmark's creation and what works with its creators — xeophon · 2026-11-20
- DTMR tackles timing mismatch when legged robots imitate human motion — KyleMorgenstein · 2026-11-20