Nature paper: AutoDNA lab runs DNA experiments from synthesis to sequencing, with humans only restocking supplies
The system orchestrates more than 20 instrument types across over 11,000 steps and matched expert-optimised enzymatic DNA synthesis yields.
Why it matters
Autonomous labs so far have mostly worked in chemistry and materials, on single goals with simple instruments. The AutoDNA authors argue that treating AI as an add-on tool is the limit, and instead co-designed models, experiments and instruments. Across four applications the system ran protocols 3.2 times faster when executing them concurrently, completed a full DNA data-storage cycle with exact recovery, and identified a clinically detected KRAS G13D mutation. What to watch is whether other groups can reproduce it and offer it as the 'science-as-a-service' the paper proposes.
Key facts
- Published in Nature on 4 November 2026
- More than 20 instrument types orchestrated
- More than 11,000 steps
- Concurrent protocols ran 3.2 times faster
- Identified a KRAS G13D mutation