Stanford's Kundaje Lab ports Illumina's PromoterAI to PyTorch, validates TERT promoter mutation scores
anshulkundaje · x · 2026-09-29
Kundaje Lab released promoterai-torch, a PyTorch port of Illumina's PromoterAI for prioritizing functional promoter variants. Running models in the same environment as benchmarking/interpretability pipelines reduces friction and standardizes the field. The post validates on TERT's recurrent C228T/C250T mutations: ensemble scores of 0.74 and 0.85, both past the paper's ±0.5 strong-effect threshold, consistent with gain-of-function telomerase reactivation.
More from Research
- Google Releases ATLAS: 15M Real Human-AI Interactions Across 150 Countries — soumitrashukla9 · 2026-09-30
- Six frontier models benchmarked across 34 capabilities in nine computer vision areas — ducha_aiki · 2026-09-30
- ETH's LP-ACRL Trains ANYmal to Run 2.5 m/s Over Rough Terrain — ChongZzZhang · 2026-09-30
- RecursiveMAS: Agent swarms share latent thoughts like looped Transformers, cutting tokens by 75% — james_y_zou · 2026-09-30
- Simplex Diffusion Models proposed to fix information collapse in discrete diffusion — msalbergo · 2026-09-30
- Physicist-turned-AI-safety-researcher publishes resource guide: SPAR, MATS, PIBBSS and more — IAmTimNguyen · 2026-09-30