gLM2 genomic language model designs new-to-nature biosynthetic enzymes
anshulkundaje · x · 2026-09-23
A preprint on bioRxiv from Tatta Bio, UC Berkeley, and LBNL shows that evolutionary sequence information learned by the genomic language model gLM2 can design complex multi-domain enzymes with chemistry beyond their natural repertoire. Key points:
- The team generated new-to-nature biosynthetic assembly lines using gLM2
- New therapeutics: over half of FDA-approved small-molecule drugs in four decades derive from natural products; redirecting biosynthetic assembly lines could greatly expand accessible chemical space
- Sustainable chemistry: biology-based routes could replace petrochemical feedstocks for materials and commodity chemicals
Authors include Jay Keasling and other synthetic biology heavyweights, marking a notable AI-for-Science advance in enzyme engineering.
More from AGI Musings
- AI floods French administrative courts as case filings surge 44% in two years — IgorCarron · 2026-09-23
- A UChicago professor used AI to help produce 200+ papers this year — soumitrashukla9 · 2026-09-23
- Alibaba CEO Wu Yongming: machine thinking could hit 1000x human levels; vibe coding is just the lightbulb — prdeepakbabu · 2026-09-23
- Polymarket puts only 19% odds on OpenAI declaring AGI before 2027 despite GPT-6 Astra — Polymarket · 2026-09-23
- Jeffrey Katzenberg on AI creativity: a demo that felt like Luxo Jr. in 1986 — elonmusk · 2026-09-23
- Hinton calls building uncontrollable AI "very stupid"; Altman says he's confident it's safe — DavidSKrueger · 2026-09-23