The DOI matches its Crossref record; 1 of the 1 references it cites were found in a registry or an index.
Highly accurate ab initio gene annotation with ANNEVO
Open accessPreprint — published as 10.1038/s41592-026-03036-7 · Paper page
Abstract
Accurate gene annotation is essential for deciphering the mapping from genomic sequences to their functional roles. However, current methods struggle to model complex gene transmission patterns and evolutionary variations in gene lengths. Here, we introduce ANNEVO, a mixture of experts (MoE)-based genomic language model that directly models distal sequence dependencies and joint evolutionary relationships from diverse genomes, enabling precise ab initio gene annotation. Through extensive benchmarking on 566 species, we demonstrate that ANNEVO significantly outperforms existing ab initio methods and achieves performance comparable to state-of-the-art annotation pipelines. Furthermore, ANNEVO’s independence from external evidence allows it to deliver more complete annotations than reference annotations for a broad range of species while correcting errors within them. These advancements will significantly advance genome sequence interpretation and provide a framework capable of integrating evolutionary insights.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Highly accurate protein structure prediction with AlphaFold
Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03819-2 · Cited by 43,764
Flagged first, then by citations
Research works printed but not linked: 24
- Lewin, H.A. et al., Vol. 119 e2115635118 (National Academy of Sciences, 2022).
- Harrison, P.W. et al. Ensembl 2024. Nucleic acids research 52, D891-D899 (2024). · DOI · Paper page
- Thibaud-Nissen, F., Souvorov, A., Murphy, T., DiCuccio, M. & Kitts, P. Eukaryotic genome annotation pipeline. The NCBI handbook 2 (2013).
- Holt, C. & Yandell, M. MAKER2: an annotation pipeline and genome-database management tool for second-generation genome projects. BMC bioinformatics 12, 1–14 (2011). · DOI · Paper page
- Brůna, T., Lomsadze, A. & Borodovsky, M. GeneMark-ETP significantly improves the accuracy of automatic annotation of large eukaryotic genomes. Genome Research 34, 757–768 (2024). · DOI
- Vosseberg, J. et al. The emerging view on the origin and early evolution of eukaryotic cells. Nature 633, 295–305 (2024). · DOI
- SÆTRE, G.P. & Saether, S.A. Ecology and genetics of speciation in Ficedula flycatchers. Molecular Ecology 19, 1091–1106 (2010). · DOI
- Parkin, I.A. et al. Transcriptome and methylome profiling reveals relics of genome dominance in the mesopolyploid Brassica oleracea. Genome biology 15, 1–18 (2014). · DOI
- Chen, L., DeVries, A.L. & Cheng, C.-H.C. Evolution of antifreeze glycoprotein gene from a trypsinogen gene in Antarctic notothenioid fish. Proceedings of the National Academy of Sciences 94, 3811–3816 (1997).
- Stanke, M. et al. AUGUSTUS: ab initio prediction of alternative transcripts. Nucleic acids research 34, W435-W439 (2006). · DOI · Paper page
- Stiehler, F. et al. Helixer: cross-species gene annotation of large eukaryotic genomes using deep learning. Bioinformatics 36, 5291–5298 (2021). · DOI
- Holst, F. et al. Helixer–de novo prediction of primary eukaryotic gene models combining deep learning and a hidden Markov model. BioRxiv, 2023.2002. 2006.527280 (2023).
- Zhou, Y. et al. Gene fusion as an important mechanism to generate new genes in the genus Oryza. Genome biology 23, 130 (2022). · DOI
- Bang, M.-L. et al. The complete gene sequence of titin, expression of an unusual ≈ 700-kDa titin isoform, and its interaction with obscurin identify a novel Z-line to I-band linking system. Circulation research 89, 1065–1072 (2001). · DOI
- Vaswani, A. Attention is all you need. Advances in Neural Information Processing Systems (2017).
- Viterbi, A. Error bounds for convolutional codes and an asymptotically optimum decoding algorithm. IEEE transactions on Information Theory 13, 260–269 (1967). · DOI
- Dalla-Torre, H. et al. Nucleotide Transformer: building and evaluating robust foundation models for human genomics. Nature Methods, 1–11 (2024).
- Brixi, G. et al. Genome modeling and design across all domains of life with Evo 2. bioRxiv, 2025.2002. 2018.638918 (2025).
- Nguyen, E. et al. Sequence modeling and design from molecular to genome scale with Evo. Science 386, eado9336 (2024). · DOI · Paper page
- Wang, S. et al. De novo and somatic structural variant discovery with SVision-pro. Nature Biotechnology, 1–5 (2024).
- Lin, J. et al. SVision: a deep learning approach to resolve complex structural variants. Nature methods 19, 1230–1233 (2022). · DOI
- Seppey, M., Manni, M. & Zdobnov, E.M. BUSCO: assessing genome assembly and annotation completeness. Gene prediction: methods and protocols, 227–245 (2019).
- He, K., Zhang, X., Ren, S. & Sun, J. in Proceedings of the IEEE conference on computer vision and pattern recognition 770–778 (2016).
- Ma, J. et al. in Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining 1930–1939 (2018).
Metadata
- Authors
- Ye, K., Zhang, P., Xu, T., Wang, S., Yang, X., Sun, P., Jia, P., Wang, B., Bush, S., & Ning, Z.
- Affiliations
- First Affiliated Hospital of Xi'an Jiaotong University (CN)
- Xi'an Jiaotong University (CN)
- Second Affiliated Hospital of Xi'an Jiaotong University (CN)
- Wellcome Sanger Institute (GB)
- Corresponding author
- Ye, Kai
- ORCID
- Venue
- Nature Methods
- Published
- 2025-04-24
- Volume
- 23
- Issue
- 4
- Pages
- 740-748
- ISSN
- 1548-7091 (ISSN-L)
- OpenAlex
- W7135034101
- Cited by
- 34
- References
- 1
- Citations by year
- 2026: 30 · 2025: 4
- FWCI
- 62.39
- Type
- journal article
- Language
- en
- Open access
- green
- doi.org
- Locations
- Nature Methods (publishedVersion)
- Research Square (acceptedVersion, cc-by)
- PubMed (publishedVersion)
- Funders
- National Natural Science Foundation of China (32125009, 323B2015, 32400509, 62172325, 32422019 and 62172325, 62302386, 32430017)
- Natural Science Foundation of Shaanxi Province (2024JC-JCQN-28)
- National Key Research and Development Program of China
- Fundamental Research Funds for the Central Universities
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Fractal and DNA sequence analysis
- Keywords
- mixture of experts, genomic language models, horizontal gene transfer, vertical inheritance, phylogenetic relationships, deep learning for genomics
- MeSH terms
- Algorithms, Animals, Humans, Models, Genetic, Phylogeny, Software, Genome, Evolution, Molecular, Computational Biology, Genomics, Molecular Sequence Annotation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
