Ernie-Switcher: RNA foundation model-guided design of programmable RNA devices in a navigable landscape

Abstract

Motivation: Recent advances in biological foundation models have enabled new opportunities for data-driven molecular engineering. However, designing programmable RNA devices remains challenging because of the vast, nonlinear sequence-structure-function landscape. Here, we present Ernie-Switcher, a deep learning framework built upon the RNA foundation model (RNAErnie), comprising three coupled components: (i) integration of RNAErnie-derived sequence-phenotype representations for realistic design generation and filtering, (ii) characterization of RNA landscape navigability as a theoretical basis for sequence optimization, and (iii) incorporation of biophysical constraints to maintain optimization trajectories within biologically plausible bounds.

Results: We validated Ernie-Switcher using RNA toehold switches. Theoretically, the observed lower bound on threshold navigability (86.4%) of the RNA landscape supports the theoretical feasibility of RNA switch sequence optimization. Experimentally, the optimized toehold switches achieved ON/OFF expression ratios of up to 35-fold in Escherichia coli. Collectively, RNAErnie-guided sequence optimization provides a scalable strategy for engineering high-performance RNA devices. To facilitate broad adoption, we deployed Ernie-Switcher as a freely accessible web platform.

Availability and implementation: The Python code and pre-trained models are available at https://github.com/zhutiblah/Ernie-Switcher and are archived on Zenodo at https://doi.org/10.5281/zenodo.21859277. The RNA toehold switch design platform is available at https://ernie-switcher.bioinformatics-syn.org/.

Optically addressable and programmable spins in DNA
K. Meng, C. Einholz, L. Krueger +2
bioRxiv14 Sep 2026PreprintAll rights reservedFull textBiophysicsbioRxiv
Programmable DNA-peptide nanostructures for multivalent regulation of intracellular signalling
M. Zacharopoulou, Z. Cassidy, S. Qin +6
bioRxiv3 Sep 2026PreprintAll rights reservedFull textBiophysicsbioRxiv
Programmable and selective silencing of oncogenic fusion transcripts by CRISPR-Cas13b via breakpoint sequence recognition
Wenxin Hu (胡文欣), Honglin Chen (陈竑霖), Steven He +18
Science Advances16 Oct 2026Peer-reviewedCC BY-NCFull text1 citedMolecular biology & gene regulationCrossrefEurope PMC
Programmable design of functional proteins from natural language
Fengyuan Dai, Shiyang You, Yudian Zhu +30
Nature21 Oct 2026Peer-reviewedLicence not statedFull text20 citedMolecular biology & gene regulationOpenAlex
Engineering gRNA for CRISPR biosensing: chemical, structural, and programmable designs for ultrasensitive diagnostics
Wenyu Fu, Rui Sang, Ruier Xue +4
TrAC Trends in Analytical Chemistry5 Jun 2026Peer-reviewedCC BY6 citedMolecular biology & gene regulationOpenAlex

Cite

BibTeX RIS

BibTeX

@article{lei2026ernieswitcher,
  title = {{Ernie-Switcher: RNA foundation model-guided design of programmable RNA devices in a navigable landscape}},
  author = {Xin Lei and Runting Li and Yuzhong Liu and Jiesheng Huang and Yicheng Yang and Ruyuan Li and Shimin Liu and Pengxia Wang and Huaiguang Jiang and Wei Xiong},
  journal = {Bioinformatics},
  year = {2026},
  month = oct,
  doi = {10.1093/bioinformatics/btag774},
  url = {https://doi.org/10.1093/bioinformatics/btag774},
  note = {CC BY 4.0},
}

RIS

TY  - JOUR
TI  - Ernie-Switcher: RNA foundation model-guided design of programmable RNA devices in a navigable landscape
AU  - Xin Lei
AU  - Runting Li
AU  - Yuzhong Liu
AU  - Jiesheng Huang
AU  - Yicheng Yang
AU  - Ruyuan Li
AU  - Shimin Liu
AU  - Pengxia Wang
AU  - Huaiguang Jiang
AU  - Wei Xiong
PY  - 2026
DA  - 2026/10/26
JO  - Bioinformatics
DO  - 10.1093/bioinformatics/btag774
UR  - https://doi.org/10.1093/bioinformatics/btag774
AB  - Motivation: Recent advances in biological foundation models have enabled new opportunities for data-driven molecular engineering. However, designing programmable RNA devices remains challenging because of the vast, nonlinear sequence-structure-function landscape. Here, we present Ernie-Switcher, a deep learning framework built upon the RNA foundation model (RNAErnie), comprising three coupled components: (i) integration of RNAErnie-derived sequence-phenotype representations for realistic design generation and filtering, (ii) characterization of RNA landscape navigability as a theoretical basis for sequence optimization, and (iii) incorporation of biophysical constraints to maintain optimization trajectories within biologically plausible bounds.  Results: We validated Ernie-Switcher using RNA toehold switches. Theoretically, the observed lower bound on threshold navigability (86.4%) of the RNA landscape supports the theoretical feasibility of RNA switch sequence optimization. Experimentally, the optimized toehold switches achieved ON/OFF expression ratios of up to 35-fold in Escherichia coli. Collectively, RNAErnie-guided sequence optimization provides a scalable strategy for engineering high-performance RNA devices. To facilitate broad adoption, we deployed Ernie-Switcher as a freely accessible web platform.  Availability and implementation: The Python code and pre-trained models are available at https://github.com/zhutiblah/Ernie-Switcher and are archived on Zenodo at https://doi.org/10.5281/zenodo.21859277. The RNA toehold switch design platform is available at https://ernie-switcher.bioinformatics-syn.org/.
SN  - 1367-4803
SN  - 1367-4811
ER  -

Abstract

Related

Cite