{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,10]],"date-time":"2026-04-10T05:23:22Z","timestamp":1775798602133,"version":"3.50.1"},"reference-count":23,"publisher":"Oxford University Press (OUP)","issue":"9","license":[{"start":{"date-parts":[[2025,9,4]],"date-time":"2025-09-04T00:00:00Z","timestamp":1756944000000},"content-version":"vor","delay-in-days":3,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100010663","name":"European Research Council","doi-asserted-by":"publisher","award":["101118521"],"award-info":[{"award-number":["101118521"]}],"id":[{"id":"10.13039\/100010663","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,9,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Motivation<\/jats:title>\n                    <jats:p>Accurately predicting how DNA sequence drives gene regulation and how genetic variants alter gene expression is a central challenge in genomics. Borzoi, which models over ten thousand genomic assays including RNA-seq coverage from over half a megabase of sequence context alone promises to become an important foundation model in regulatory genomics, both for massively annotating variants and for further model development. However, the currently used relative positional encodings limit Borzoi\u2019s computational efficiency.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>We present Flashzoi, an enhanced Borzoi model that leverages rotary positional encodings and FlashAttention-2. This achieves over 3-fold faster training and inference and up to 2.4-fold reduced memory usage, while maintaining or improving accuracy in modeling various genomic assays including RNA-seq coverage, predicting variant effects, and enhancer-promoter linking. Flashzoi\u2019s improved efficiency facilitates large-scale genomic analyses and opens avenues for exploring more complex regulatory mechanisms and modeling.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Availability and implementation<\/jats:title>\n                    <jats:p>The Flashzoi model architecture is part of the MIT-licensed borzoi-pytorch package, can be found at https:\/\/github.com\/johahi\/borzoi-pytorch and installed via pip. Model weights for all four Flashzoi and Borzoi replicates are available at https:\/\/huggingface.co\/johahi under the MIT license. The code has been archived at https:\/\/zenodo.org\/records\/15669913.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btaf467","type":"journal-article","created":{"date-parts":[[2025,9,3]],"date-time":"2025-09-03T11:50:19Z","timestamp":1756900219000},"source":"Crossref","is-referenced-by-count":5,"title":["Flashzoi: an enhanced Borzoi for accelerated genomic analysis"],"prefix":"10.1093","volume":"41","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8260-032X","authenticated-orcid":false,"given":"Johannes C","family":"Hingerl","sequence":"first","affiliation":[{"name":"School of Computation, Information and Technology, Technical University of Munich , Garching, 85748,","place":["Germany"]},{"name":"Munich Center for Machine Learning , Munich, 80333,","place":["Germany"]}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7570-7877","authenticated-orcid":false,"given":"Alexander","family":"Karollus","sequence":"additional","affiliation":[{"name":"School of Computation, Information and Technology, Technical University of Munich , Garching, 85748,","place":["Germany"]},{"name":"Munich Center for Machine Learning , Munich, 80333,","place":["Germany"]}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8924-8365","authenticated-orcid":false,"given":"Julien","family":"Gagneur","sequence":"additional","affiliation":[{"name":"School of Computation, Information and Technology, Technical University of Munich , Garching, 85748,","place":["Germany"]},{"name":"Munich Center for Machine Learning , Munich, 80333,","place":["Germany"]},{"name":"Institute of Human Genetics, School of Medicine, Technical University of Munich , Munich, 81675,","place":["Germany"]},{"name":"Computational Health Center, Helmholtz Center Munich , Neuherberg, 85764,","place":["Germany"]}]}],"member":"286","published-online":{"date-parts":[[2025,9,4]]},"reference":[{"key":"2025092403220452000_btaf467-B1","author":"Ainslie","year":"2023"},{"key":"2025092403220452000_btaf467-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":"2025092403220452000_btaf467-B3","doi-asserted-by":"crossref","first-page":"354","DOI":"10.1038\/s41588-021-00782-6","article-title":"Base-resolution models of transcription-factor binding reveal soft motif syntax","volume":"53","author":"Avsec","year":"2021","journal-title":"Nat Genet"},{"key":"2025092403220452000_btaf467-B5","author":"Dai","year":"2019"},{"key":"2025092403220452000_btaf467-B6","author":"Dao","year":"2024"},{"key":"2025092403220452000_btaf467-B7","author":"Dong","year":"2024"},{"key":"2025092403220452000_btaf467-B8","author":"Drusinsky","year":"2024"},{"key":"2025092403220452000_btaf467-B9","doi-asserted-by":"crossref","first-page":"57","DOI":"10.1038\/nature11247","article-title":"An integrated encyclopedia of DNA elements in the human genome","volume":"489","author":"Dunham","year":"2012","journal-title":"Nature"},{"key":"2025092403220452000_btaf467-B10","doi-asserted-by":"crossref","first-page":"D942","DOI":"10.1093\/nar\/gkac1071","article-title":"GENCODE: reference annotation for the human and mouse genomes in 2023","volume":"51","author":"Frankish","year":"2023","journal-title":"Nucleic Acids Res"},{"key":"2025092403220452000_btaf467-B11","author":"Gschwind","year":"2023"},{"key":"2025092403220452000_btaf467-B12","author":"Hingerl","year":"2024"},{"key":"2025092403220452000_btaf467-B13","doi-asserted-by":"crossref","first-page":"56","DOI":"10.1186\/s13059-023-02899-9","article-title":"Current sequence-based models capture gene expression determinants in promoters but mostly ignore distal enhancers","volume":"24","author":"Karollus","year":"2023","journal-title":"Genome Biol"},{"key":"2025092403220452000_btaf467-B14","doi-asserted-by":"crossref","first-page":"1290","DOI":"10.1038\/s41588-021-00924-w","article-title":"A compendium of uniformly processed human gene expression and splicing quantitative trait loci","volume":"53","author":"Kerimov","year":"2021","journal-title":"Nat Genet"},{"key":"2025092403220452000_btaf467-B15","author":"Lal","year":"2024"},{"key":"2025092403220452000_btaf467-B16","doi-asserted-by":"crossref","first-page":"949","DOI":"10.1038\/s41588-024-02053-6","article-title":"Predicting RNA-seq coverage from DNA sequence as a unifying model of gene regulation","volume":"57","author":"Linder","year":"2025","journal-title":"Nat Genet"},{"key":"2025092403220452000_btaf467-B4"},{"key":"2025092403220452000_btaf467-B17","author":"Loshchilov","year":"2019"},{"key":"2025092403220452000_btaf467-B18","first-page":"3349","author":"Martyn"},{"key":"2025092403220452000_btaf467-B19","author":"Paszke"},{"key":"2025092403220452000_btaf467-B20","doi-asserted-by":"crossref","first-page":"2060","DOI":"10.1038\/s41588-023-01524-6","article-title":"Benchmarking of deep neural networks for predicting personal gene expression from DNA sequence highlights shortcomings","volume":"55","author":"Sasse","year":"2023","journal-title":"Nat Genet"},{"key":"2025092403220452000_btaf467-B21","author":"Schwessinger","year":"2023"},{"key":"2025092403220452000_btaf467-B22","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"},{"key":"2025092403220452000_btaf467-B24","author":"Vaswani"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btaf467\/64211188\/btaf467.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/41\/9\/btaf467\/64211188\/btaf467.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/41\/9\/btaf467\/64211188\/btaf467.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,24]],"date-time":"2025-09-24T07:22:11Z","timestamp":1758698531000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/doi\/10.1093\/bioinformatics\/btaf467\/8248080"}},"subtitle":[],"editor":[{"given":"Jianlin","family":"Cheng","sequence":"additional","affiliation":[]}],"short-title":[],"issued":{"date-parts":[[2025,9,1]]},"references-count":23,"journal-issue":{"issue":"9","published-print":{"date-parts":[[2025,9,1]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btaf467","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2024.12.18.629121","asserted-by":"object"}]},"ISSN":["1367-4803","1367-4811"],"issn-type":[{"value":"1367-4803","type":"print"},{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2025,9]]},"published":{"date-parts":[[2025,9,1]]},"article-number":"btaf467"}}