{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T15:56:16Z","timestamp":1791388576342,"version":"4.3.3"},"reference-count":27,"publisher":"Oxford University Press (OUP)","issue":"9","license":[{"start":{"date-parts":[[2022,3,3]],"date-time":"2022-03-03T00:00:00Z","timestamp":1646265600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/journals\/pages\/open_access\/funder_policies\/chorus\/standard_publication_model"}],"funder":[{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["5U01HG009431"],"award-info":[{"award-number":["5U01HG009431"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["1DP2GM123485"],"award-info":[{"award-number":["1DP2GM123485"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2022,4,28]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Motivation<\/jats:title>\n                    <jats:p>Deep-learning models, such as convolutional neural networks, are able to accurately map biological sequences to associated functional readouts and properties by learning predictive de novo representations. In silico saturation mutagenesis (ISM) is a popular feature attribution technique for inferring contributions of all characters in an input sequence to the model\u2019s predicted output. The main drawback of ISM is its runtime, as it involves multiple forward propagations of all possible mutations of each character in the input sequence through the trained model to predict the effects on the output.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>We present fastISM, an algorithm that speeds up ISM by a factor of over 10\u00d7 for commonly used convolutional neural network architectures. fastISM is based on the observations that the majority of computation in ISM is spent in convolutional layers, and a single mutation only disrupts a limited region of intermediate layers, rendering most computation redundant. fastISM reduces the gap between backpropagation-based feature attribution methods and ISM. It far surpasses the runtime of backpropagation-based methods on multi-output architectures, making it feasible to run ISM on a large number of sequences.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Availability and implementation<\/jats:title>\n                    <jats:p>An easy-to-use Keras\/TensorFlow 2 implementation of fastISM is available at https:\/\/github.com\/kundajelab\/fastISM. fastISM can be installed using pip install fastism. A hands-on tutorial can be found at https:\/\/colab.research.google.com\/github\/kundajelab\/fastISM\/blob\/master\/notebooks\/colab\/DeepSEA.ipynb.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Supplementary information<\/jats:title>\n                    <jats:p>Supplementary data are available at Bioinformatics online.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btac135","type":"journal-article","created":{"date-parts":[[2022,3,1]],"date-time":"2022-03-01T07:12:29Z","timestamp":1646118749000},"page":"2397-2403","source":"Crossref","is-referenced-by-count":30,"title":["fastISM: performant\n                    <i>in silico<\/i>\n                    saturation mutagenesis for convolutional neural networks"],"prefix":"10.1093","volume":"38","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6216-2457","authenticated-orcid":false,"given":"Surag","family":"Nair","sequence":"first","affiliation":[{"name":"Department of Computer Science, Stanford University , Stanford, CA 94305, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Avanti","family":"Shrikumar","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Stanford University , Stanford, CA 94305, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jacob","family":"Schreiber","sequence":"additional","affiliation":[{"name":"Department of Genetics, Stanford University , Stanford, CA 94305, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3084-2287","authenticated-orcid":false,"given":"Anshul","family":"Kundaje","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Stanford University , Stanford, CA 94305, USA"},{"name":"Department of Genetics, Stanford University , Stanford, CA 94305, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2022,3,3]]},"reference":[{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"831","DOI":"10.1038\/nbt.3300","article-title":"Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning","volume":"33","author":"Alipanahi","year":"2015","journal-title":"Nat. Biotechnol"},{"key":"2023041402565583800_","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":"2023041402565583800_","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":"2023041402565583800_","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":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"1111","DOI":"10.1038\/s41592-020-0958-x","article-title":"Predicting 3D genome folding from DNA sequence with Akita","volume":"17","author":"Fudenberg","year":"2020","journal-title":"Nat. Methods"},{"key":"2023041402565583800_","first-page":"178","article-title":"DeeperBind: enhancing prediction of sequence specificities of DNA binding proteins","volume":"2016","author":"Hassanzadeh","year":"2016","journal-title":"Proceedings (IEEE Int. Conf. Bioinformatics Biomed.)"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Comput"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1016\/j.cell.2018.12.015","article-title":"Predicting splicing from primary sequence with deep learning","volume":"176","author":"Jaganathan","year":"2019","journal-title":"Cell"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"149","DOI":"10.1186\/s13059-020-02055-7","article-title":"Enhanced Integrated Gradients: improving interpretability of deep learning m1qodels using splicing codes as a case study","volume":"21","author":"Jha","year":"2020","journal-title":"Genome Biol"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"990","DOI":"10.1101\/gr.200535.115","article-title":"Basset: learning the regulatory code of the accessible genome with deep convolutional neural networks","volume":"26","author":"Kelley","year":"2016","journal-title":"Genome Res"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"739","DOI":"10.1101\/gr.227819.117","article-title":"Sequential regulatory activity prediction across chromosomes with convolutional neural networks","volume":"28","author":"Kelley","year":"2018","journal-title":"Genome Res"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"258","DOI":"10.1038\/s42256-020-00291-x","article-title":"Improving representations of genomic sequence motifs in convolutional networks with exponential activations","volume":"3","author":"Koo","year":"2021","journal-title":"Nat. Mach. Intell"},{"key":"2023041402565583800_","article-title":"Inferring sequence-structure preferences of RNA-binding proteins with convolutional residual networks","author":"Koo","year":"2018","journal-title":"BioRxiv"},{"key":"2023041402565583800_","author":"Lundberg","year":"2017"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"1173","DOI":"10.1038\/nbt.1589","article-title":"High-resolution analysis of DNA regulatory elements by synthetic saturation mutagenesis","volume":"27","author":"Patwardhan","year":"2009","journal-title":"Nat. Biotechnol"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"e107","DOI":"10.1093\/nar\/gkw226","article-title":"DanQ: a hybrid convolutional and recurrent deep neural network for quantifying the function of DNA sequences","volume":"44","author":"Quang","year":"2016","journal-title":"Nucleic Acids Res"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"15270","DOI":"10.1038\/s41598-018-33321-1","article-title":"Recurrent neural network for predicting transcription factor binding sites","volume":"8","author":"Shen","year":"2018","journal-title":"Sci. Rep"},{"key":"2023041402565583800_","first-page":"3145","author":"Shrikumar","year":"2017"},{"key":"2023041402565583800_","author":"Shrikumar","year":"2018"},{"key":"2023041402565583800_","author":"Simonyan","year":"2014"},{"key":"2023041402565583800_","first-page":"3319","author":"Sundararajan","year":"2017"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"1301","DOI":"10.1016\/j.csbj.2019.12.011","article-title":"Deep learning methods in protein structure prediction","volume":"18","author":"Torrisi","year":"2020","journal-title":"Comput. Struct. Biotechnol. J"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"i269","DOI":"10.1093\/bioinformatics\/btz339","article-title":"Comprehensive evaluation of deep learning architectures for prediction of DNA\/RNA sequence binding specificities","volume":"35","author":"Trabelsi","year":"2019","journal-title":"Bioinformatics"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"e51503","DOI":"10.7554\/eLife.51503","article-title":"Deep learning models predict regulatory variants in pancreatic islets and refine type 2 diabetes association signals","volume":"9","author":"Wesolowska-Andersen","year":"2020","journal-title":"Elife"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"119","DOI":"10.1016\/j.isci.2019.09.018","article-title":"Deep learning implicitly handles tissue specific phenomena to predict tumor DNA accessibility and immune activity","volume":"20","author":"Wnuk","year":"2019","journal-title":"iScience"},{"key":"2023041402565583800_","doi-asserted-by":"crossref","first-page":"931","DOI":"10.1038\/nmeth.3547","article-title":"Predicting effects of noncoding variants with deep learning-based sequence model","volume":"12","author":"Zhou","year":"2015","journal-title":"Nat. Methods"},{"key":"2023041402565583800_","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"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btac135\/42782147\/btac135.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/38\/9\/2397\/49874221\/btac135.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/38\/9\/2397\/49874221\/btac135.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,11,18]],"date-time":"2023-11-18T02:29:04Z","timestamp":1700274544000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/38\/9\/2397\/6541633"}},"subtitle":[],"editor":[{"given":"Pier Luigi","family":"Martelli","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2022,3,3]]},"references-count":27,"journal-issue":{"issue":"9","published-print":{"date-parts":[[2022,4,28]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btac135","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2020.10.13.337147","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":[[2022,5,1]]},"published":{"date-parts":[[2022,3,3]]}}}