Genome language models uncover new class of reverse-transcriptase mechanisms
BrianHie · x · 2026-09-23
Brian Hie's team (David Li, Garyk Brixi, with the Michael Fischbach lab) released Minerva, an approach using genome language models for systematic biological discovery.
- AI accelerates biological discovery, not just biological design
- Genome language models systematically mine unannotated sequences to surface previously unknown genetic elements
- They reveal a new class of reverse-transcriptase mechanisms: UG27 reverse-transcriptase systems encode variable arrays of diverse ncRNAs with a shared structure, each templating a short DNA hairpin
A notable AI-for-Science result showing language models can drive real biological hypothesis generation and element discovery.
More from Research
- New paper reframes score distillation as distribution matching, explains SDS mode collapse — burny_tech · 2026-09-23
- Cambridge publishes open-access 458-page book on nonparametric and high-dim stats — FrnkNlsn · 2026-09-23
- Reinforce-Ada: adaptive sampling for RLVR recovers lost signals, 2x faster convergence — burny_tech · 2026-09-23
- Paper: Transformers do have world models — failures trace to feature interference — burny_tech · 2026-09-23
- Paper suggests transformers can't learn loop-closure in world modeling — burny_tech · 2026-09-23
- Epoch AI: The plunging price of thought — intelligence keeps getting radically cheaper — Proper_Actuary2907 · 2026-09-23