CDS-BART: a BART-based foundation model for mRNA sequence analysis.
- 1MOGAM Institute for Biomedical Research, Seoul, 06730, Republic of Korea.
Abstract
Summary
Recent advancements in artificial intelligence (AI) have led to the development of foundation models that interpret mRNA as a language. Notable examples include CodonBERT, HydraRNA, Evo 2, and Helix-mRNA. These models demonstrate significant potential as powerful tools for mRNA research. However, to the best of our knowledge, there is currently no publicly available AI model that is both easy to use and capable of analysing mRNA sequences up to about 4 kb, a length scale typical of many therapeutic mRNAs, including those encapsulated within lipid nanoparticles. Thus, we propose CDS-BART, a user-friendly, open-source tool that integrates SentencePiece subword tokenization with the denoising sequence-to-sequence training of Bidirectional and Auto-Regressive Transformers (BART). CDS-BART was pre-trained on mRNA data from nine taxonomic groups provided by the NCBI RefSeq database. This comprehensive pre-training, coupled with BART's denoising capability, supports learning of coding sequence (CDS)-level sequence regularities that are useful for downstream mRNA property prediction. Thus, CDS-BART can ultimately deliver robust performance across a wide range of mRNA prediction tasks.Availability and implementation
CDS-BART is released under the MIT License. Latest code is available via GitHub at https://github.com/mogam-ai/CDS-BART and archived on Zenodo (DOI: 10.5281/zenodo.21502775).
Related
- Translation initiation by the Kozak mRNA sequence is based on a conformational readout on the ribosome
- Joint protein, mRNA and DNA sequence design and optimization with nucleotide-level Potts models
- Sequence- and Base-Selective Modification of mRNA within and nearby Premature Termination Codon and Its Effect on the Translation in a Human Cell-Free Translation System
- Advancing Targeted Nanoparticle mRNA Delivery Using Machine Learning
- Analytical evaluation of eukaryotic cell-free translation (CFT) systems to assess mRNA translatability
BibTeX
@article{jadamba2026cdsbart,
title = {{CDS-BART: a BART-based foundation model for mRNA sequence analysis.}},
author = {Erkhembayar Jadamba and Sang-Heon Lee and Sungho Lee and Hyekyoung Lee and Jinhee Hong and Hyunjin Shin},
journal = {Bioinformatics advances},
year = {2026},
month = sep,
doi = {10.1093/bioadv/vbag259},
url = {https://europepmc.org/article/MED/42927291},
note = {CC BY},
}
RIS
TY - JOUR
TI - CDS-BART: a BART-based foundation model for mRNA sequence analysis.
AU - Erkhembayar Jadamba
AU - Sang-Heon Lee
AU - Sungho Lee
AU - Hyekyoung Lee
AU - Jinhee Hong
AU - Hyunjin Shin
PY - 2026
DA - 2026/09/15
JO - Bioinformatics advances
DO - 10.1093/bioadv/vbag259
UR - https://europepmc.org/article/MED/42927291
AB - Summary Recent advancements in artificial intelligence (AI) have led to the development of foundation models that interpret mRNA as a language. Notable examples include CodonBERT, HydraRNA, Evo 2, and Helix-mRNA. These models demonstrate significant potential as powerful tools for mRNA research. However, to the best of our knowledge, there is currently no publicly available AI model that is both easy to use and capable of analysing mRNA sequences up to about 4 kb, a length scale typical of many therapeutic mRNAs, including those encapsulated within lipid nanoparticles. Thus, we propose CDS-BART, a user-friendly, open-source tool that integrates SentencePiece subword tokenization with the denoising sequence-to-sequence training of Bidirectional and Auto-Regressive Transformers (BART). CDS-BART was pre-trained on mRNA data from nine taxonomic groups provided by the NCBI RefSeq database. This comprehensive pre-training, coupled with BART's denoising capability, supports learning of coding sequence (CDS)-level sequence regularities that are useful for downstream mRNA property prediction. Thus, CDS-BART can ultimately deliver robust performance across a wide range of mRNA prediction tasks.Availability and implementation CDS-BART is released under the MIT License. Latest code is available via GitHub at https://github.com/mogam-ai/CDS-BART and archived on Zenodo (DOI: 10.5281/zenodo.21502775).
SN - 2635-0041
ER -