Protein Language Models Can Now Dream Up Enzymes Nature Never Made
Protein language models can write enzymes unlike anything in nature. We review the peer-reviewed evidence on what works, what fails and what comes next.
Editor's picks
See all
Plastic-Eating Enzymes Designed by AI: How Close Are We to Real Recycling?
AI-designed plastic-eating enzymes now break down PET in hours in the lab, yet the first industrial plant is on hold. A balanced look at what the evidence shows.
Inside the AI Labs Where Digital Scientists Argue, Test and Discover
Sewage as an Early Warning System: Can AI Spot the Next Outbreak?
AI Spots Rare Diseases Doctors Miss. Should Yours Be Using It?
AI Molecular Clocks Promise Earlier Cancer Detection. Do They Deliver?
AI
See all
Virology
See all
Recent posts
See all
AI Can Now Design Antibodies From Scratch. Here's the Catch
Protein language models and structure predictors can now design antibodies from scratch. We look at the hit rates, the evidence, the limits and the safety questions.
Who Regulates an AI That Designs a Virus? The Global Policy Gap
AI can now write working virus genomes. Who is responsible for checking them? A plain-English tour of AI biosecurity rules in the US, EU, UK and beyond.
Mpox Is Still Spreading. Can AI Get Ahead of the Virus This Time?
Mpox is still circulating, and new clades are mixing. We look at what genomic surveillance has revealed, where machine learning genuinely helps, and where the evidence stops.
Can Scientists Trust AI Research Assistants? What the Evidence Shows
AI research assistants can speed up literature searches, but fabricated citations and patchy recall remain real risks. Here is what 2026 studies say about which tools hold up.
AI Agents Are Learning to Run the Lab. This Week They Got a Boost
OpenAI's DevDay 2026 turned long-running autonomous agents into everyday infrastructure. Here is what that means for the slow, expensive work of automating scientific research, and where the honest limits still sit.
Medical AI Just Hit 90% Accuracy. But Why Can't Hospitals Trust It?
A new Nature Medicine study shows on-premise medical AI agents can diagnose at near-cloud accuracy. The bigger story is why the field still can't agree on how to know when to trust them.
Biology
See allAI Molecular Clocks Promise Earlier Cancer Detection. Do They Deliver?
Epigenetic clocks and machine learning are being pointed at early cancer detection. We separate what the 2026 evidence shows from what the headlines imply, from breast cancer risk to the NHS-Galleri trial.
