Designing RNAs with Language Models
Milan Gautam, Ning Dai, Tianshuo Zhou, Bowen Xie, David Mathews, Liang Huang · 2026
Read Designing RNAs with Language Models by Milan Gautam, Ning Dai, Tianshuo Zhou, Bowen Xie, David Mathews, Liang Huang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-12470.txt.
Description
RNA design asks for a sequence that folds into a target secondary structure, a task with huge sequence space and many competing folds. Milan Gautam led this work with five co-authors, reframing design as conditional sequence generation rather than per-instance optimisation. A reusable neural approximator, built as an autoregressive language model, maps target structures directly to sequences.
Training starts supervised on random structure-sequence pairs, then reinforcement learning optimises end-to-end metrics, with methods to pick a small subset for reinforcement learning that greatly improve efficiency and quality. Across four data sets, the approach beats state-of-the-art systems on key metrics such as Boltzmann probability, while running 1.7 times faster. Conditional language-model generation thus offers a scalable, task-agnostic alternative to per-instance search for RNA design.
Catalog record
| CREATOR | Milan Gautam, Ning Dai, Tianshuo Zhou, Bowen Xie, David Mathews, Liang Huang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-12470