structRFM
structRFM is a rna structure & design tool whose code is published by heqin-zhu on GitHub, first public in 2026, described in Nature (2026). Open structure-guided RNA foundation model trained on 21 million sequence–structure pairs; uses pair-matching masked LM (MUSES ensemble) for secondary structure, zero-shot homology, Zfold 3D and function tasks.
| Category | RNA structure & design |
|---|---|
| Developer (GitHub) | heqin-zhu |
| Task | RNA secondary/tertiary structure and function prediction from sequence |
| Interface | Python (pip install structRFM), Hugging Face transformers |
| GitHub | heqin-zhu/structRFM |
| Released | 2026-11-04 (source) |
| Paper | https://doi.org/10.1038/s41586-026-11210-2 · Nature (2026) |
| Code license | MIT |
| Weights license | weights via GitHub Release and Hugging Face heqin-zhu/structRFM with no separate weights license stated |
| What this allows | Permissive (commercial OK) Use, modify and ship it inside closed-source commercial products; keep the license text and copyright notices. What do these mean? |
| License source | github.com/heqin-zhu/structRFM/blob/master/LICENSE · checked 2026-11-05 |
| Pricing | Open source |
| Tags | #foundation-model#rna-structure#secondary-structure#language-model |
| Last checked | · links verified |
Can I use structRFM commercially?
Yes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices.
- Code: MIT
- Weights: weights via GitHub Release and Hugging Face heqin-zhu/structRFM with no separate weights license stated
- Source: github.com/heqin-zhu/structRFM/blob/master/LICENSE · license checked 2026-11-05
What the license types mean · not legal advice.
Specs
| Task | RNA secondary/tertiary structure and function prediction from sequence |
|---|---|
| Structure level (2D / 3D) | — |
| Input | — |
| GitHub stars | 58 |
| Interface | Python (pip install structRFM), Hugging Face transformers |
Compare with…
Pick 1–3 other RNA structure & design tools to compare side by side with structRFM.
Related
Standard RNA secondary structure prediction and design package (RNAfold, RNAinverse).
RNA foundation model pretrained on 23M non-coding RNAs for structure and function tasks.
Geometric deep learning for 3D RNA inverse design (sequence design from backbone).