Deep Learning for RNA Design Makes Science Cover, AI Matches Expert Humans on Pseudoknots
rishabh16_ · x · 2026-09-04
Chaitanya Joshi announced that gRNAde, a deep learning system for RNA design that began as a side project during his CS PhD, has been published in Science and chosen as the cover article.
Key result: on the Eterna citizen-science platform, AI methods matched expert human players at inventing sequences that fold into RNA pseudoknots — complex structures where loops pair with distant parts of the same molecule.
This opens a route to automatically designing RNA therapeutics, sensors, and catalysts from scratch, capping a three-year journey from new ML modeling to experimental validation.
More from Research
- YC-affiliated team offers $500k and free YAM arms for dexterous manipulation benchmarks — ycombinator · 2026-09-04
- MITRA team lands two WMT2026 papers: largest Sanskrit-English corpus and 9B translation model — SebastianNehrd2 · 2026-09-04
- Reusing reasoning traces as pre-context state lifts long-context accuracy in 26 of 27 tests — burny_tech · 2026-09-04
- Running Full Scientific Experiments on Robots with Claude: New Protocol in Under 10 Minutes — yawnxyz · 2026-09-04
- 753B model 'thinks', 4B model writes: latent-space handoff claimed to be 20x faster — burny_tech · 2026-09-04
- TailSFT: Microsoft Paper Shows SFT Can Wreck RL Starting Points, Gains up to 4% pass@1 — rohanpaul_ai · 2026-09-04