Stanford unveils ENCODE GRAMMAR: seq2func models across 3,865 datasets decode regulatory DNA
anshulkundaje · x · 2026-10-15
At a Berkeley BioML seminar, Stanford PhD candidate Chang Yun presents ENCODE GRAMMAR, one of the largest collections of regulatory DNA sequence models to date.
- Built from 3,865 experimental datasets spanning multiple layers of gene regulation, a seq2func model atlas
- Accompanying interpretation frameworks identify the sequence features and rules driving predictions: which bases matter, how arrangements cause cell-type-specific behavior, how variants alter activity
- Demonstrates integrating predictions and interpretations across models and modalities to decode regulatory sequence logic at scale
Related event: Stanford Unveils ENCODE GRAMMAR: 3,865 Datasets Decoding Regulatory DNA(2 posts)→
More from Research
- Player-1: RL turns VLM agents into GTA V players with zero game-specific training — yshan2u · 2026-10-15
- Von Neumann's trick: how to get a fair flip from an unfair coin — doodlestein · 2026-10-15
- AI-assisted preprint improves kissing number bounds in 19 and 21 dimensions — felpix_ · 2026-10-15
- Gracia's new world model targets Real2Sim2Real and Physical AI — jonstephens85 · 2026-10-15
- Niantic Spatial built VPS on Pokémon Go imagery; a dev already prototyped it solo — jonstephens85 · 2026-10-15
- SFT learns arithmetic mixture, OPD learns geometric mean: why distillation forgets differently — QuanquanGu · 2026-10-15