{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,10]],"date-time":"2026-11-10T05:55:54Z","timestamp":1794290154083,"version":"4.9.0"},"reference-count":47,"publisher":"Springer Science and Business Media LLC","issue":"10","license":[{"start":{"date-parts":[[2021,10,1]],"date-time":"2021-10-01T00:00:00Z","timestamp":1633046400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,10,4]],"date-time":"2021-10-04T00:00:00Z","timestamp":1633305600000},"content-version":"vor","delay-in-days":3,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Nat Methods"],"published-print":{"date-parts":[[2021,10]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    How noncoding DNA determines gene expression in different cell types is a major unsolved problem, and critical downstream applications in human genetics depend on improved solutions. Here, we report substantially improved gene expression prediction accuracy from DNA sequences through the use of a deep learning architecture, called Enformer, that is able to integrate information from long-range interactions (up to 100\u2009kb away) in the genome. This improvement yielded more accurate variant effect predictions on gene expression for both natural genetic variants and saturation mutagenesis measured by massively parallel reporter assays. Furthermore, Enformer learned to predict enhancer\u2013promoter interactions directly from the DNA sequence competitively with methods that take direct experimental data as input. We expect that these advances will enable more effective fine-mapping of human disease associations and provide a framework to interpret\n                    <jats:italic>cis<\/jats:italic>\n                    -regulatory evolution.\n                  <\/jats:p>","DOI":"10.1038\/s41592-021-01252-x","type":"journal-article","created":{"date-parts":[[2021,10,4]],"date-time":"2021-10-04T13:38:04Z","timestamp":1633354684000},"page":"1196-1203","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":1528,"title":["Effective gene expression prediction from sequence by integrating long-range interactions"],"prefix":"10.1038","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7790-8936","authenticated-orcid":false,"given":"\u017diga","family":"Avsec","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vikram","family":"Agarwal","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Daniel","family":"Visentin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Joseph R.","family":"Ledsam","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Agnieszka","family":"Grabska-Barwinska","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kyle R.","family":"Taylor","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yannis","family":"Assael","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"John","family":"Jumper","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7466-7997","authenticated-orcid":false,"given":"Pushmeet","family":"Kohli","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7782-3548","authenticated-orcid":false,"given":"David R.","family":"Kelley","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2021,10,4]]},"reference":[{"key":"1252_CR1","doi-asserted-by":"publisher","first-page":"1171","DOI":"10.1038\/s41588-018-0160-6","volume":"50","author":"J Zhou","year":"2018","unstructured":"Zhou, J. et al. Deep learning sequence-based ab initio prediction of variant effects on expression and disease risk. Nat. Genet. 50, 1171\u20131179 (2018).","journal-title":"Nat. Genet."},{"key":"1252_CR2","doi-asserted-by":"publisher","first-page":"e1008050","DOI":"10.1371\/journal.pcbi.1008050","volume":"16","author":"DR Kelley","year":"2020","unstructured":"Kelley, D. R. Cross-species regulatory sequence activity prediction. PLoS Comput. Biol. 16, e1008050 (2020).","journal-title":"PLoS Comput. Biol."},{"key":"1252_CR3","doi-asserted-by":"publisher","first-page":"739","DOI":"10.1101\/gr.227819.117","volume":"28","author":"DR Kelley","year":"2018","unstructured":"Kelley, D. R. et al. Sequential regulatory activity prediction across chromosomes with convolutional neural networks. Genome Res. 28, 739\u2013750 (2018).","journal-title":"Genome Res."},{"key":"1252_CR4","doi-asserted-by":"publisher","first-page":"107663","DOI":"10.1016\/j.celrep.2020.107663","volume":"31","author":"V Agarwal","year":"2020","unstructured":"Agarwal, V. & Shendure, J. Predicting mRNA abundance directly from genomic sequence using deep convolutional neural networks. Cell Rep. 31, 107663 (2020).","journal-title":"Cell Rep."},{"key":"1252_CR5","doi-asserted-by":"publisher","first-page":"292","DOI":"10.1038\/s41576-019-0209-0","volume":"21","author":"M Gasperini","year":"2020","unstructured":"Gasperini, M., Tome, J. M. & Shendure, J. Towards a comprehensive catalogue of validated and target-linked human enhancers. Nat. Rev. Genet. 21, 292\u2013310 (2020).","journal-title":"Nat. Rev. Genet."},{"key":"1252_CR6","unstructured":"Vaswani, A. et al. Attention is all you need. in Advances in Neural Information Processing Systems 5998\u20136008 (2017)."},{"key":"1252_CR7","unstructured":"Brown, T. B. et al. Language models are few-shot learners. in Advances in Neural Information Processing Systems (2020)."},{"key":"1252_CR8","doi-asserted-by":"publisher","unstructured":"Ji, Y., Zhou, Z., Liu, H. & Davuluri, R. V. DNABERT: pre-trained bidirectional encoder representations from transformers model for DNA-language in genome. Bioinformatics https:\/\/doi.org\/10.1093\/bioinformatics\/btab083 (2021).","DOI":"10.1093\/bioinformatics\/btab083"},{"key":"1252_CR9","doi-asserted-by":"publisher","first-page":"377","DOI":"10.1016\/j.cell.2018.11.029","volume":"176","author":"M Gasperini","year":"2019","unstructured":"Gasperini, M. et al. A genome-wide framework for mapping gene regulation via cellular genetic screens. Cell 176, 377\u2013390 (2019).","journal-title":"Cell"},{"key":"1252_CR10","doi-asserted-by":"publisher","first-page":"462","DOI":"10.1038\/nature13182","volume":"507","author":"FANTOM Consortium and the RIKEN PMI and CLST (DGT).","year":"2014","unstructured":"FANTOM Consortium and the RIKEN PMI and CLST (DGT).A promoter-level mammalian expression atlas. Nature 507, 462\u2013470 (2014).","journal-title":"Nature"},{"key":"1252_CR11","doi-asserted-by":"publisher","first-page":"144","DOI":"10.1038\/nrm3949","volume":"16","author":"S Heinz","year":"2015","unstructured":"Heinz, S., Romanoski, C. E., Benner, C. & Glass, C. K. The selection and function of cell type-specific enhancers. Nat. Rev. Mol. Cell Biol. 16, 144\u2013154 (2015).","journal-title":"Nat. Rev. Mol. Cell Biol."},{"key":"1252_CR12","unstructured":"Shrikumar, A., Greenside, P. & Kundaje, A. Learning important features through propagating activation differences. in International Conference on Machine Learning 3145\u20133153 (PMLR, 2017)."},{"key":"1252_CR13","doi-asserted-by":"publisher","first-page":"1664","DOI":"10.1038\/s41588-019-0538-0","volume":"51","author":"CP Fulco","year":"2019","unstructured":"Fulco, C. P. et al. Activity-by-contact model of enhancer-promoter regulation from thousands of CRISPR perturbations. Nat. Genet. 51, 1664\u20131669 (2019).","journal-title":"Nat. Genet."},{"key":"1252_CR14","doi-asserted-by":"publisher","first-page":"389","DOI":"10.1038\/s41576-019-0122-6","volume":"20","author":"G Eraslan","year":"2019","unstructured":"Eraslan, G., Avsec, \u017d., Gagneur, J. & Theis, F. J. Deep learning: new computational modelling techniques for genomics. Nat. Rev. Genet. 20, 389\u2013403 (2019).","journal-title":"Nat. Rev. Genet."},{"key":"1252_CR15","doi-asserted-by":"publisher","unstructured":"Avsec, \u017d. et al. Base-resolution models of transcription-factor binding reveal soft motif syntax. Nat. Genet. https:\/\/doi.org\/10.1038\/s41588-021-00782-6 (2021).","DOI":"10.1038\/s41588-021-00782-6"},{"key":"1252_CR16","doi-asserted-by":"publisher","first-page":"699","DOI":"10.1038\/s41586-020-2493-4","volume":"583","author":"ENCODE Project Consortium","year":"2020","unstructured":"ENCODE Project Consortiumet al. Expanded encyclopaedias of DNA elements in the human and mouse genomes. Nature 583, 699\u2013710 (2020).","journal-title":"Nature"},{"key":"1252_CR17","doi-asserted-by":"publisher","first-page":"2205","DOI":"10.1093\/bioinformatics\/btw203","volume":"32","author":"M Ghandi","year":"2016","unstructured":"Ghandi, M. et al. gkmSVM: an R package for gapped-kmer SVM. Bioinformatics 32, 2205\u20132207 (2016).","journal-title":"Bioinformatics"},{"key":"1252_CR18","doi-asserted-by":"publisher","first-page":"931","DOI":"10.1038\/nmeth.3547","volume":"12","author":"J Zhou","year":"2015","unstructured":"Zhou, J. & Troyanskaya, O. G. Predicting effects of noncoding variants with deep learning\u2013based sequence model. Nat. Methods 12, 931\u2013934 (2015).","journal-title":"Nat. Methods"},{"key":"1252_CR19","doi-asserted-by":"publisher","first-page":"990","DOI":"10.1101\/gr.200535.115","volume":"26","author":"DR Kelley","year":"2016","unstructured":"Kelley, D. R., Snoek, J. & Rinn, J. L. Basset: learning the regulatory code of the accessible genome with deep convolutional neural networks. Genome Res. 26, 990\u2013999 (2016).","journal-title":"Genome Res."},{"key":"1252_CR20","doi-asserted-by":"publisher","first-page":"1318","DOI":"10.1126\/science.aaz1776","volume":"369","author":"TG Consortium","year":"2020","unstructured":"Consortium, T. G., The GTEx Consortium. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318\u20131330 (2020).","journal-title":"Science"},{"key":"1252_CR21","doi-asserted-by":"publisher","first-page":"1483","DOI":"10.1038\/s41588-018-0196-7","volume":"50","author":"YA Reshef","year":"2018","unstructured":"Reshef, Y. A. et al. Detecting genome-wide directional effects of transcription factor binding on polygenic disease risk. Nat. Genet. 50, 1483\u20131493 (2018).","journal-title":"Nat. Genet."},{"key":"1252_CR22","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-021-23134-8","volume":"12","author":"QS Wang","year":"2021","unstructured":"Wang, Q. S. et al. Leveraging supervised learning for functionally informed fine-mapping of cis-eQTLs identifies an additional 20,913 putative causal eQTLs. Nat. Commun. 12, 3394 (2021).","journal-title":"Nat. Commun."},{"key":"1252_CR23","doi-asserted-by":"publisher","first-page":"1273","DOI":"10.1111\/rssb.12388","volume":"82","author":"G Wang","year":"2020","unstructured":"Wang, G., Sarkar, A., Carbonetto, P. & Stephens, M. A simple new approach to variable selection in regression, with application to genetic fine mapping. J. R. Stat. Soc. 82, 1273\u20131300 (2020).","journal-title":"J. R. Stat. Soc."},{"key":"1252_CR24","doi-asserted-by":"publisher","first-page":"1355","DOI":"10.1038\/s41588-020-00735-5","volume":"52","author":"O Weissbrod","year":"2020","unstructured":"Weissbrod, O. et al. Functionally informed fine-mapping and polygenic localization of complex trait heritability. Nat. Genet. 52, 1355\u20131363 (2020).","journal-title":"Nat. Genet."},{"key":"1252_CR25","doi-asserted-by":"crossref","unstructured":"Kircher, M., Xiong, C., Martin, B. & Schubach, M. Saturation mutagenesis of twenty disease-associated regulatory elements at single base-pair resolution. Nature 10, 3583 (2019).","DOI":"10.1038\/s41467-019-11526-w"},{"key":"1252_CR26","doi-asserted-by":"publisher","first-page":"1280","DOI":"10.1002\/humu.23797","volume":"40","author":"D Shigaki","year":"2019","unstructured":"Shigaki, D. et al. Integration of multiple epigenomic marks improves prediction of variant impact in saturation mutagenesis reporter assay. Hum. Mutat. 40, 1280\u20131291 (2019).","journal-title":"Hum. Mutat."},{"key":"1252_CR27","doi-asserted-by":"publisher","first-page":"955","DOI":"10.1038\/ng.3331","volume":"47","author":"D Lee","year":"2015","unstructured":"Lee, D. et al. A method to predict the impact of regulatory variants from DNA sequence. Nat. Genet. 47, 955\u2013961 (2015).","journal-title":"Nat. Genet."},{"key":"1252_CR28","doi-asserted-by":"publisher","first-page":"D886","DOI":"10.1093\/nar\/gky1016","volume":"47","author":"P Rentzsch","year":"2019","unstructured":"Rentzsch, P., Witten, D., Cooper, G. M., Shendure, J. & Kircher, M. CADD: predicting the deleteriousness of variants throughout the human genome. Nucleic Acids Res. 47, D886\u2013D894 (2019).","journal-title":"Nucleic Acids Res."},{"key":"1252_CR29","doi-asserted-by":"publisher","first-page":"831","DOI":"10.1038\/nbt.3300","volume":"33","author":"B Alipanahi","year":"2015","unstructured":"Alipanahi, B., Delong, A., Weirauch, M. T. & Frey, B. J. Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning. Nat. Biotechnol. 33, 831\u2013838 (2015).","journal-title":"Nat. Biotechnol."},{"key":"1252_CR30","doi-asserted-by":"publisher","first-page":"554","DOI":"10.1016\/j.cell.2015.01.006","volume":"160","author":"D Villar","year":"2015","unstructured":"Villar, D. et al. Enhancer evolution across 20 mammalian species. Cell 160, 554 (2015).","journal-title":"Cell"},{"key":"1252_CR31","doi-asserted-by":"crossref","unstructured":"Linder, J. et al. A generative neural network for maximizing fitness and diversity of synthetic DNA and protein sequences. Cell Syst. 11, 49\u201362 (2020).","DOI":"10.1016\/j.cels.2020.05.007"},{"key":"1252_CR32","doi-asserted-by":"publisher","first-page":"1111","DOI":"10.1038\/s41592-020-0958-x","volume":"17","author":"G Fudenberg","year":"2020","unstructured":"Fudenberg, G., Kelley, D. R. & Pollard, K. S. Predicting 3D genome folding from DNA sequence with Akita. Nat. Methods 17, 1111\u20131117 (2020).","journal-title":"Nat. Methods"},{"key":"1252_CR33","doi-asserted-by":"publisher","first-page":"1118","DOI":"10.1038\/s41592-020-0960-3","volume":"17","author":"R Schwessinger","year":"2020","unstructured":"Schwessinger, R. et al. DeepC: predicting 3D genome folding using megabase-scale transfer learning. Nat. Methods 17, 1118\u20131124 (2020).","journal-title":"Nat. Methods"},{"key":"1252_CR34","doi-asserted-by":"publisher","DOI":"10.1186\/s13059-020-01977-6","volume":"21","author":"J Schreiber","year":"2020","unstructured":"Schreiber, J., Durham, T., Bilmes, J. & Noble, W. S. Avocado: a multi-scale deep tensor factorization method learns a latent representation of the human epigenome. Genome Biol. 21, 81 (2020).","journal-title":"Genome Biol."},{"key":"1252_CR35","doi-asserted-by":"publisher","first-page":"i108","DOI":"10.1093\/bioinformatics\/btz352","volume":"35","author":"S Nair","year":"2019","unstructured":"Nair, S., Kim, D. S., Perricone, J. & Kundaje, A. Integrating regulatory DNA sequence and gene expression to predict genome-wide chromatin accessibility across cellular contexts. Bioinformatics 35, i108\u2013i116 (2019).","journal-title":"Bioinformatics"},{"key":"1252_CR36","unstructured":"Tay, Y., Dehghani, M., Bahri, D. & Metzler, D. Efficient transformers: a survey. Preprint at https:\/\/arxiv.org\/abs\/2009.06732 (2020)."},{"key":"1252_CR37","doi-asserted-by":"publisher","first-page":"769","DOI":"10.1038\/s41588-020-0652-z","volume":"52","author":"F Richter","year":"2020","unstructured":"Richter, F. et al. Genomic analyses implicate noncoding de novo variants in congenital heart disease. Nat. Genet. 52, 769\u2013777 (2020).","journal-title":"Nat. Genet."},{"key":"1252_CR38","doi-asserted-by":"publisher","first-page":"973","DOI":"10.1038\/s41588-019-0420-0","volume":"51","author":"J Zhou","year":"2019","unstructured":"Zhou, J. et al. Whole-genome deep-learning analysis identifies contribution of noncoding mutations to autism risk. Nat. Genet. 51, 973\u2013980 (2019).","journal-title":"Nat. Genet."},{"key":"1252_CR39","doi-asserted-by":"publisher","DOI":"10.1186\/gb-2007-8-2-r24","volume":"8","author":"S Gupta","year":"2007","unstructured":"Gupta, S., Stamatoyannopoulos, J. A., Bailey, T. L. & Noble, W. Quantifying similarity between motifs. Genome Biology 8, R24 (2007).","journal-title":"Genome Biology"},{"key":"1252_CR40","doi-asserted-by":"crossref","unstructured":"Shaw, P., Uszkoreit, J. & Vaswani, A. Self-attention with relative position representations. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers) 464\u2013468 (2018).","DOI":"10.18653\/v1\/N18-2074"},{"key":"1252_CR41","doi-asserted-by":"crossref","unstructured":"Dai, Z. et al. Transformer-XL: Attentive language models beyond a fixed-length context. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics 2978\u20132988 (2019).","DOI":"10.18653\/v1\/P19-1285"},{"key":"1252_CR42","doi-asserted-by":"publisher","first-page":"996","DOI":"10.1101\/gr.229102","volume":"12","author":"WJ Kent","year":"2002","unstructured":"Kent, W. J. The Human Genome Browser at UCSC. Genome Research 12, 996\u20131006 (2002).","journal-title":"Genome Research"},{"key":"1252_CR43","unstructured":"Reynolds, M. et al. Open sourcing Sonnet \u2014 a new library for constructing neural networks. https:\/\/deepmind.com\/blog\/open-sourcing-sonnet (2017)."},{"key":"1252_CR44","first-page":"2825","volume":"12","author":"F Pedregosa","year":"2011","unstructured":"Pedregosa, F. et al. Scikit-learn: machine learning in Python. J. Mach. Learn. Res. 12, 2825\u20132830 (2011).","journal-title":"J. Mach. Learn. Res."},{"key":"1252_CR45","doi-asserted-by":"publisher","first-page":"1083","DOI":"10.1038\/s41592-020-0965-y","volume":"17","author":"JC Klein","year":"2020","unstructured":"Klein, J. C. et al. A systematic evaluation of the design and context dependencies of massively parallel reporter assays. Nat. Methods 17, 1083\u20131091 (2020).","journal-title":"Nat. Methods"},{"key":"1252_CR46","doi-asserted-by":"publisher","unstructured":"Avsec, \u017diga et al. Enformer (Version 3.0) (Zenodo, 2021); https:\/\/doi.org\/10.5281\/zenodo.5098375","DOI":"10.5281\/zenodo.5098375"},{"key":"1252_CR47","doi-asserted-by":"publisher","first-page":"592","DOI":"10.1038\/s41587-019-0140-0","volume":"37","author":"\u017d Avsec","year":"2019","unstructured":"Avsec, \u017d. et al. The Kipoi repository accelerates community exchange and reuse of predictive models for genomics. Nat. Biotechnol. 37, 592\u2013600 (2019).","journal-title":"Nat. Biotechnol."}],"container-title":["Nature Methods"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41592-021-01252-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41592-021-01252-x","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41592-021-01252-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,10,20]],"date-time":"2021-10-20T11:28:23Z","timestamp":1634729303000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41592-021-01252-x"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,10]]},"references-count":47,"journal-issue":{"issue":"10","published-print":{"date-parts":[[2021,10]]}},"alternative-id":["1252"],"URL":"https:\/\/doi.org\/10.1038\/s41592-021-01252-x","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2021.04.07.438649","asserted-by":"object"}]},"ISSN":["1548-7091","1548-7105"],"issn-type":[{"value":"1548-7091","type":"print"},{"value":"1548-7105","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,10]]},"assertion":[{"value":"17 February 2021","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"27 July 2021","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"4 October 2021","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"\u017d.A., A.G-B., K.R.T., Y.A., J.J., and P.K. are employed by DeepMind. V.A., and D.R.K. are employed by Calico Life Sciences. J.R.L. is employed by Google. The remaining authors declare no competing interests.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}]}}