{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,19]],"date-time":"2026-10-19T08:48:08Z","timestamp":1792399688741,"version":"4.4.0"},"reference-count":22,"publisher":"Oxford University Press (OUP)","issue":"4","license":[{"start":{"date-parts":[[2026,10,19]],"date-time":"2026-10-19T00:00:00Z","timestamp":1792368000000},"content-version":"vor","delay-in-days":21,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1936791"],"award-info":[{"award-number":["1936791"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["2107108"],"award-info":[{"award-number":["2107108"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000104","name":"National Aeronautics and Space Administration","doi-asserted-by":"publisher","award":["80NSSC20M0097"],"award-info":[{"award-number":["80NSSC20M0097"]}],"id":[{"id":"10.13039\/100000104","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100005758","name":"Pennsylvania Space Grant Consortium","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100005758","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,9,28]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Semi-supervised learning (SSL) is a natural fit for genomics, where unlabeled sequences are abundant but high-quality labels are scarce due to costly curation. However, most SSL pipelines rely on generic perturbations that are not tailored to biological constraints, and it remains unclear which augmentation designs yield reliable gains for DNA sequence classification. In this application note, we present an empirical study of\u00a0genome-aware weak\u2013strong augmentation recipes for pseudo-label-based SSL, focusing on FixMatch and FlexMatch with a GROVER DNA language model backbone. Across two classification benchmarks\u2014antimicrobial resistance (AMR) drug-class prediction from AMR gene sequences and multi-class oncovirus family classification from simulated 150\u2009bp tumor-like reads\u2014we systematically evaluate 25 weak\/strong augmentation pairs constructed from biologically motivated operators, including reverse-complement, codon-level back-translation, mild masking\/substitution (NN), high-rate substitution, and length-preserving insertion\u2013deletion (indel). We find consistent trends: (i) moderate, sequence-preserving perturbations yield the best accuracy\u2013stability trade-off, (ii) excessive masking or aggressive mutation increases variance and degrades performance, and (iii) short-segment masking (NN) for weak augmentation followed by an indel rate (around 5%) for strong augmentation emerge as a robust strategy to learn from unlabeled genomic data. We release implementations and recommended defaults to facilitate reproducible SSL for genomic sequence modeling under limited-label regimes.<\/jats:p>","DOI":"10.1093\/nargab\/lqag123","type":"journal-article","created":{"date-parts":[[2026,10,19]],"date-time":"2026-10-19T08:32:57Z","timestamp":1792398777000},"source":"Crossref","is-referenced-by-count":0,"title":["DNA augmentations for semi-supervised learning in genomic sequence classification"],"prefix":"10.1093","volume":"8","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3862-4006","authenticated-orcid":false,"given":"Hyunwoo","family":"Yoo","sequence":"first","affiliation":[{"name":"Department of Electrical and Computer Engineering, Drexel University , 3141 Chestnut St, Philadelphia, PA 19104 ,","place":["United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1763-5750","authenticated-orcid":false,"given":"Gail L","family":"Rosen","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Drexel University , 3141 Chestnut St, Philadelphia, PA 19104 ,","place":["United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2026,10,19]]},"reference":[{"key":"2026101904325311900_B1","doi-asserted-by":"publisher","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":"2026101904325311900_B2","article-title":"FixMatch: simplifying semi-supervised learning with consistency and confidence","volume-title":"Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS 2020)","author":"Sohn","year":"2020"},{"key":"2026101904325311900_B3","article-title":"FlexMatch: boosting semi-supervised learning with curriculum pseudo labeling","volume-title":"Proceedings of the 35th Conference on Neural Information Processing Systems (NeurIPS 2021)","author":"Zhang","year":"2021"},{"key":"2026101904325311900_B4","doi-asserted-by":"publisher","first-page":"911","DOI":"10.1038\/s42256-024-00872-0","article-title":"DNA language model GROVER learns sequence context in the human genome","volume":"6","author":"Sanabria","year":"2024","journal-title":"Nat Mach Intell"},{"key":"2026101904325311900_B5","doi-asserted-by":"publisher","first-page":"287","DOI":"10.1038\/s41592-024-02523-z","article-title":"The Nucleotide Transformer: building and evaluating robust foundation models for human genomics","volume":"22","author":"Dalla-Torre","year":"2025","journal-title":"Nat Methods"},{"key":"2026101904325311900_B6","article-title":"DNABERT-2: efficient foundation model and benchmark for multi-species genome","author":"Zhou"},{"key":"2026101904325311900_B7","doi-asserted-by":"publisher","first-page":"e1013360","DOI":"10.1371\/journal.pcbi.1013360","article-title":"NextVir: enabling classification of tumor-causing viruses with genomic foundation models","volume":"21","author":"Robertson","year":"2025","journal-title":"PLoS Comput Biol"},{"key":"2026101904325311900_B8","article-title":"Predicting anti-microbial resistance using large language models","author":"Yoo"},{"key":"2026101904325311900_B9","doi-asserted-by":"crossref","first-page":"263","DOI":"10.18653\/v1\/2025.bionlp-1.23","article-title":"Enhancing antimicrobial drug resistance classification by integrating sequence-based and text-based representations","volume-title":"Proceedings of the 24th Workshop on Biomedical Language Processing.","author":"Yoo","year":"2025"},{"key":"2026101904325311900_B10","doi-asserted-by":"publisher","first-page":"186","DOI":"10.1186\/s12859-023-05303-2","article-title":"Semi-supervised learning improves regulatory sequence prediction with unlabeled sequences","volume":"24","author":"Mourad","year":"2023","journal-title":"BMC Bioinformatics"},{"key":"2026101904325311900_B11","doi-asserted-by":"publisher","first-page":"bbae560","DOI":"10.1093\/bib\/bbae560","article-title":"Semi-supervised learning with pseudo-labeling compares favorably with large language models for regulatory sequence prediction","volume":"25","author":"Phan","year":"2024","journal-title":"Brief Bioinform"},{"key":"2026101904325311900_B12","first-page":"6256","article-title":"Unsupervised data augmentation for consistency training","volume-title":"Advances in Neural Information Processing Systems 33 (NeurIPS 2020)","author":"Xie","year":"2020"},{"key":"2026101904325311900_B13","first-page":"18613","article-title":"RandAugment: practical automated data augmentation with a reduced search space","volume-title":"Advances in Neural Information Processing Systems 33 (NeurIPS 2020)","author":"Cubuk","year":"2020"},{"key":"2026101904325311900_B14","doi-asserted-by":"publisher","first-page":"170","DOI":"10.1186\/s12859-024-05787-6","article-title":"Assessing the reliability of point mutation as data augmentation for deep learning with genomic data","volume":"25","author":"Lee","year":"2024","journal-title":"BMC Bioinformatics"},{"key":"2026101904325311900_B15","doi-asserted-by":"publisher","first-page":"btae190","DOI":"10.1093\/bioinformatics\/btae190","article-title":"Improving the performance of supervised deep learning for regulatory genomics using phylogenetic augmentation","volume":"40","author":"Duncan","year":"2024","journal-title":"Bioinformatics"},{"key":"2026101904325311900_B16","doi-asserted-by":"publisher","first-page":"105","DOI":"10.1186\/s13059-023-02941-w","article-title":"EvoAug: improving generalization and interpretability of genomic deep neural networks with evolution-inspired data augmentations","volume":"24","author":"Lee","year":"2023","journal-title":"Genome Biol"},{"key":"2026101904325311900_B17","first-page":"1","article-title":"Towards a better understanding of reverse-complement equivariance for deep learning models in genomics","volume-title":"Proceedings of the 16th Machine Learning in Computational Biology meeting","author":"Zhou","year":"2022"},{"key":"2026101904325311900_B18","doi-asserted-by":"publisher","first-page":"vbac094","DOI":"10.1093\/bioadv\/vbac094","article-title":"Nucleotide augmentation for machine learning-guided protein engineering","volume":"3","author":"Minot","year":"2022","journal-title":"Bioinform Adv"},{"key":"2026101904325311900_B19","doi-asserted-by":"publisher","first-page":"25","DOI":"10.1186\/s12863-023-01123-8","article-title":"Genomic benchmarks: a collection of datasets for genomic sequence classification","volume":"24","author":"Gre\u0161ov\u00e1","year":"2023","journal-title":"BMC Genom Data"},{"key":"2026101904325311900_B20","doi-asserted-by":"publisher","first-page":"D561","DOI":"10.1093\/nar\/gkz1010","article-title":"MEGARes 2.0: a database for classification of antimicrobial drug, biocide and metal resistance determinants in metagenomic sequence data","volume":"48","author":"Doster","year":"2020","journal-title":"Nucleic Acids Res"},{"key":"2026101904325311900_B21","doi-asserted-by":"publisher","first-page":"D566","DOI":"10.1093\/nar\/gkw1004","article-title":"CARD 2017: expansion and model-centric curation of the comprehensive antibiotic resistance database","volume":"45","author":"Jia","year":"2017","journal-title":"Nucleic Acids Res"},{"key":"2026101904325311900_B22","first-page":"1","article-title":"Statistical comparisons of classifiers over multiple data sets","volume":"7","author":"Dem\u0161ar","year":"2006","journal-title":"J Mach Learn Res"}],"container-title":["NAR Genomics and Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/nargab\/article-pdf\/8\/4\/lqag123\/72085387\/lqag123.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/nargab\/article-pdf\/8\/4\/lqag123\/72085387\/lqag123.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,19]],"date-time":"2026-10-19T08:32:59Z","timestamp":1792398779000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/nargab\/article\/doi\/10.1093\/nargab\/lqag123\/8937181"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,28]]},"references-count":22,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,9,28]]}},"URL":"https:\/\/doi.org\/10.1093\/nargab\/lqag123","relation":{},"ISSN":["2631-9268"],"issn-type":[{"value":"2631-9268","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,12]]},"published":{"date-parts":[[2026,9,28]]},"article-number":"lqag123"}}