{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,24]],"date-time":"2026-06-24T13:48:29Z","timestamp":1782308909961,"version":"3.54.5"},"reference-count":41,"publisher":"Oxford University Press (OUP)","issue":"6","license":[{"start":{"date-parts":[[2026,5,11]],"date-time":"2026-05-11T00:00:00Z","timestamp":1778457600000},"content-version":"vor","delay-in-days":2,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000001","name":"US National Science Foundation","doi-asserted-by":"publisher","award":["III-2232121"],"award-info":[{"award-number":["III-2232121"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"US National Institutes of Health","doi-asserted-by":"publisher","award":["R01HG012470"],"award-info":[{"award-number":["R01HG012470"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,6,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Motivation<\/jats:title>\n                    <jats:p>Genomic language models (gLMs) face a fundamental efficiency challenge: one must either maintain separate specialized models for each biological modality (DNA and RNA) or develop large multimodal architectures. Both approaches impose significant computational burdens\u2014modality-specific models require redundant infrastructure despite inherent biological connections, while multi-modal architectures demand increased parameter counts and extensive cross-modality pretraining.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>To address this limitation, we introduce CodonMoE (Adaptive Mixture of Codon Reformative Experts), a lightweight adapter that transforms DNA language models into effective RNA analyzers without RNA-specific pretraining. Our theoretical analysis establishes CodonMoE as a universal approximator at the codon level, capable of mapping arbitrary functions from codon sequences to codon-dependent RNA properties given sufficient expert capacity. Across four RNA prediction tasks spanning stability, expression, and regulation, DNA models augmented with CodonMoE significantly outperform their unmodified counterparts, with the HyenaDNA+CodonMoE series achieving state-of-the-art results using 80% fewer parameters than specialized RNA models. By maintaining sub-quadratic complexity while achieving superior performance, our approach provides a principled path toward unifying genomic language modeling, leveraging more abundant DNA data and reducing computational overhead while preserving modality-specific performance advantages.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Availability and implementation<\/jats:title>\n                    <jats:p>Source code for the method and to reproduce the results is available at https:\/\/github.com\/Kingsford-Group\/CodonMoE.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btag285","type":"journal-article","created":{"date-parts":[[2026,5,5]],"date-time":"2026-05-05T11:29:10Z","timestamp":1777980550000},"source":"Crossref","is-referenced-by-count":0,"title":["CodonMoE: DNA language models for codon-dependent mRNA prediction"],"prefix":"10.1093","volume":"42","author":[{"given":"Shiyi","family":"Du","sequence":"first","affiliation":[{"name":"Ray and Stephanie Lane Computational Biology Department, Carnegie Mellon University , PA 15213,","place":["United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Litian","family":"Liang","sequence":"additional","affiliation":[{"name":"Ray and Stephanie Lane Computational Biology Department, Carnegie Mellon University , PA 15213,","place":["United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiayi","family":"Li","sequence":"additional","affiliation":[{"name":"Ray and Stephanie Lane Computational Biology Department, Carnegie Mellon University , PA 15213,","place":["United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0118-5516","authenticated-orcid":false,"given":"Carl","family":"Kingsford","sequence":"additional","affiliation":[{"name":"Ray and Stephanie Lane Computational Biology Department, Carnegie Mellon University , PA 15213,","place":["United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2026,5,9]]},"reference":[{"key":"2026062409114156500_btag285-B1","doi-asserted-by":"crossref","first-page":"lqac012","DOI":"10.1093\/nargab\/lqac012","article-title":"Informative RNA base embedding for RNA structural alignment and clustering by deep representation learning","volume":"4","author":"Akiyama","year":"2022","journal-title":"NAR Genom Bioinform"},{"key":"2026062409114156500_btag285-B27"},{"key":"2026062409114156500_btag285-B2","doi-asserted-by":"crossref","first-page":"1196","DOI":"10.1038\/s41592-021-01252-x","article-title":"Effective gene expression prediction from sequence by integrating long-range interactions","volume":"18","author":"Avsec","year":"2021","journal-title":"Nat Methods"},{"key":"2026062409114156500_btag285-B3","first-page":"1960"},{"key":"2026062409114156500_btag285-B4","first-page":"286"},{"key":"2026062409114156500_btag285-B5","doi-asserted-by":"crossref","first-page":"654","DOI":"10.1016\/j.cels.2021.05.017","article-title":"Learning the protein language: evolution, structure, and function","volume":"12","author":"Bepler","year":"2021","journal-title":"Cell Syst"},{"key":"2026062409114156500_btag285-B6","author":"Celaj","year":"2023"},{"key":"2026062409114156500_btag285-B7","author":"Chen","year":"2022"},{"key":"2026062409114156500_btag285-B8"},{"key":"2026062409114156500_btag285-B9","first-page":"287"},{"key":"2026062409114156500_btag285-B10","first-page":"4171","year":"2019"},{"key":"2026062409114156500_btag285-B11"},{"key":"2026062409114156500_btag285-B12","doi-asserted-by":"crossref","first-page":"389","DOI":"10.1038\/s41576-019-0122-6","article-title":"Deep learning: new computational modelling techniques for genomics","volume":"20","author":"Eraslan","year":"2019","journal-title":"Nat Rev Genet"},{"key":"2026062409114156500_btag285-B13","first-page":"1"},{"key":"2026062409114156500_btag285-B14","doi-asserted-by":"crossref","first-page":"34","DOI":"10.1021\/acssynbio.8b00207","article-title":"Tuning the performance of synthetic riboswitches using machine learning","volume":"8","author":"Groher","year":"2019","journal-title":"ACS Synth Biol"},{"key":"2026062409114156500_btag285-B15"},{"key":"2026062409114156500_btag285-B16","doi-asserted-by":"crossref","first-page":"359","DOI":"10.1016\/0893-6080(89)90020-8","article-title":"Multilayer feedforward networks are universal approximators","volume":"2","author":"Hornik","year":"1989","journal-title":"Neural Networks"},{"key":"2026062409114156500_btag285-B17","doi-asserted-by":"crossref","first-page":"79","DOI":"10.1162\/neco.1991.3.1.79","article-title":"Adaptive mixtures of local experts","volume":"3","author":"Jacobs","year":"1991","journal-title":"Neural Comput"},{"key":"2026062409114156500_btag285-B18","doi-asserted-by":"crossref","first-page":"2112","DOI":"10.1093\/bioinformatics\/btab083","article-title":"DNABERT: pre-trained bidirectional encoder representations from transformers model for DNA-language in genome","volume":"37","author":"Ji","year":"2021","journal-title":"Bioinformatics"},{"key":"2026062409114156500_btag285-B19","doi-asserted-by":"crossref","first-page":"181","DOI":"10.1162\/neco.1994.6.2.181","article-title":"Hierarchical mixtures of experts and the EM algorithm","volume":"6","author":"Jordan","year":"1994","journal-title":"Neural Comput"},{"key":"2026062409114156500_btag285-B20","doi-asserted-by":"crossref","first-page":"583","DOI":"10.1038\/s41586-021-03819-2","article-title":"Highly accurate protein structure prediction with AlphaFold","volume":"596","author":"Jumper","year":"2021","journal-title":"Nature"},{"key":"2026062409114156500_btag285-B21","author":"Lepikhin","year":"2021"},{"key":"2026062409114156500_btag285-B22","doi-asserted-by":"crossref","first-page":"1536","DOI":"10.1038\/s41467-022-28776-w","article-title":"Combinatorial optimization of mRNA structure, stability, and translation for RNA-based therapeutics","volume":"13","author":"Leppek","year":"2022","journal-title":"Nat Commun"},{"key":"2026062409114156500_btag285-B23","doi-asserted-by":"crossref","first-page":"1027","DOI":"10.1101\/gr.278870.123","article-title":"CodonBERT large language model for mRNA vaccines","volume":"34","author":"Li","year":"2024","journal-title":"Genome Res"},{"key":"2026062409114156500_btag285-B24"},{"key":"2026062409114156500_btag285-B25","first-page":"43177","volume":"36"},{"key":"2026062409114156500_btag285-B26","doi-asserted-by":"crossref","first-page":"2363","DOI":"10.1093\/nar\/gkad035","article-title":"Revealing determinants of translation efficiency via whole-gene codon randomization and machine learning","volume":"51","author":"Nieuwkoop","year":"2023","journal-title":"Nucleic Acids Res"},{"key":"2026062409114156500_btag285-B28"},{"key":"2026062409114156500_btag285-B30"},{"key":"2026062409114156500_btag285-B31","first-page":"9392","year":"2024"},{"key":"2026062409114156500_btag285-B32"},{"key":"2026062409114156500_btag285-B33","doi-asserted-by":"crossref","first-page":"127063","DOI":"10.1016\/j.neucom.2023.127063","article-title":"Roformer: enhanced transformer with rotary position embedding","volume":"568","author":"Su","year":"2024","journal-title":"Neurocomputing (Amst)"},{"key":"2026062409114156500_btag285-B29"},{"key":"2026062409114156500_btag285-B34","doi-asserted-by":"crossref","first-page":"D439","DOI":"10.1093\/nar\/gkab1061","article-title":"AlphaFold protein structure database: massively expanding the structural coverage of protein-sequence space with high-accuracy models","volume":"50","author":"Varadi","year":"2022","journal-title":"Nucleic Acids Res"},{"key":"2026062409114156500_btag285-B35"},{"key":"2026062409114156500_btag285-B36","doi-asserted-by":"crossref","first-page":"852","DOI":"10.1038\/s42256-022-00534-z","article-title":"scBERT as a large-scale pretrained deep language model for cell type annotation of single-cell RNA-seq data","volume":"4","author":"Yang","year":"2022","journal-title":"Nat Mach Intell"},{"key":"2026062409114156500_btag285-B37"},{"key":"2026062409114156500_btag285-B38","first-page":"e2421738122","author":"Zhai","year":"2025"},{"key":"2026062409114156500_btag285-B39"},{"key":"2026062409114156500_btag285-B40","doi-asserted-by":"crossref","first-page":"1171","DOI":"10.1038\/s41588-018-0160-6","article-title":"Deep learning sequence-based ab initio prediction of variant effects on expression and disease risk","volume":"50","author":"Zhou","year":"2018","journal-title":"Nat Genet"},{"key":"2026062409114156500_btag285-B41"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btag285\/68263309\/btag285.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/6\/btag285\/68263309\/btag285.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/6\/btag285\/68263309\/btag285.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,24]],"date-time":"2026-06-24T13:11:49Z","timestamp":1782306709000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/doi\/10.1093\/bioinformatics\/btag285\/8675376"}},"subtitle":[],"editor":[{"given":"Inanc","family":"Birol","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,5,9]]},"references-count":41,"journal-issue":{"issue":"6","published-print":{"date-parts":[[2026,6,1]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btag285","relation":{},"ISSN":["1367-4803","1367-4811"],"issn-type":[{"value":"1367-4803","type":"print"},{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,6]]},"published":{"date-parts":[[2026,5,9]]},"article-number":"btag285"}}