De novo peptide sequencing with InstaNovo: Diffusion-powered, peptide identification for large scale proteomics experiments | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article De novo peptide sequencing with InstaNovo: Diffusion-powered, peptide identification for large scale proteomics experiments Timothy Jenkins, Kevin Eloff, Konstantinos Kalogeropoulos, Oliver Morell, and 13 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3376248/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 31 Mar, 2025 Read the published version in Nature Machine Intelligence → Version 1 posted You are reading this latest preprint version Abstract Bottom-up mass spectrometry-based proteomics is challenged by the task of identifying the peptide that generates a tandem mass spectrum. Traditional methods that rely on known peptide sequence databases are limited and may not be applicable in certain contexts. De novo peptide sequencing, which assigns peptide sequences to the spectra without prior information, is valuable for various biological applica-tions; yet, due to a lack of accuracy, it remains challenging to apply this approach in many situations. Here, we introduce InstaNovo, a transformer neural network with the ability to translate fragment ion peaks into the sequence of amino acids that make up the studied pep-tide(s). The model was trained on 28 million labelled spectra matched to 742k human peptides from the ProteomeTools project. We demon-strate that InstaNovo outperforms current state-of-the-art methods on benchmark datasets and showcase its utility in several applications. Building upon human intuition, we also introduce InstaNovo+, a multi-nomial diffusion model that further improves performance by iterative refinement of predicted sequences. Using these models, we could de novo sequence antibody-based therapeutics with unprecedented cov-erage, discover novel peptides, and detect unreported organisms in different datasets, thereby expanding the scope and detection rate of proteomics searches. Finally, we could experimentally validate tryp-tic and non-tryptic peptides with targeted proteomics, demonstrat-ing the fidelity of our predictions. Our models unlock a plethora of opportunities across different scientific domains, such as direct protein sequencing, immunopeptidomics, and exploration of the dark proteome. Biological sciences/Computational biology and bioinformatics/Proteome informatics Biological sciences/Computational biology and bioinformatics/Machine learning Full Text Additional Declarations Yes there is potential Competing Interest. KE, JVG, WW, MS, KB, and NC are employees of InstaDeep, 5 Merchant Square, London, UK. The remaining authors declare no conflicts of interest. Cite Share Download PDF Status: Published Journal Publication published 31 Mar, 2025 Read the published version in Nature Machine Intelligence → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3376248","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":296889791,"identity":"41b433fe-8a85-4a8f-99b2-ca362ec95592","order_by":0,"name":"Timothy Jenkins","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAoElEQVRIiWNgGAWjYJACA4YKZhBtTJxyHrCWM8wSpGlhYGwjRYu9dPODgo/zrOsMDjBvNiDOFpljBoYzt6VLGBxgK04gTotEgoEx77bDQC08xgeI1JL+wfjvHNK05BgYMzZAtBDpsDtnCgx7jqVLzjzMVkyc99lnt28z+FFjzc93vHmzBFFaGCQY2CCGMxOnHqyF+QHRikfBKBgFo2BkAgAjtyoSaxxIEwAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-2979-5663","institution":"Department of Biotechnology and Biomedicine, Technical University of Denmark","correspondingAuthor":true,"prefix":"","firstName":"Timothy","middleName":"","lastName":"Jenkins","suffix":""},{"id":296889792,"identity":"9360d071-45a3-4d75-bc62-df0624358345","order_by":1,"name":"Kevin Eloff","email":"","orcid":"","institution":"InstaDeep","correspondingAuthor":false,"prefix":"","firstName":"Kevin","middleName":"","lastName":"Eloff","suffix":""},{"id":296889793,"identity":"18fc0fa7-c5f5-4b7f-9ffe-8bb6ff392191","order_by":2,"name":"Konstantinos Kalogeropoulos","email":"","orcid":"https://orcid.org/0000-0003-3907-9281","institution":"Department of Biotechnology and Biomedicine, Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Konstantinos","middleName":"","lastName":"Kalogeropoulos","suffix":""},{"id":296889794,"identity":"bbc8f93c-9b80-4bbf-92fb-4c7fe1ab2c65","order_by":3,"name":"Oliver Morell","email":"","orcid":"https://orcid.org/0009-0000-8702-1792","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Oliver","middleName":"","lastName":"Morell","suffix":""},{"id":296889795,"identity":"58536f7f-030e-453a-87ca-d8e276df0223","order_by":4,"name":"Amandla Mabona","email":"","orcid":"","institution":"InstaDeep","correspondingAuthor":false,"prefix":"","firstName":"Amandla","middleName":"","lastName":"Mabona","suffix":""},{"id":296889796,"identity":"c245a790-f192-4ab8-ad8f-0e172dbe96d7","order_by":5,"name":"Jakob Berg Jespersen","email":"","orcid":"","institution":"Department of Biotechnology and Biomedicine, Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Jakob","middleName":"Berg","lastName":"Jespersen","suffix":""},{"id":296889797,"identity":"c2c709b3-496f-4f63-9228-e10a5a3661a8","order_by":6,"name":"Wesley Williams","email":"","orcid":"","institution":"InstaDeep","correspondingAuthor":false,"prefix":"","firstName":"Wesley","middleName":"","lastName":"Williams","suffix":""},{"id":296889798,"identity":"c561d01b-078c-4b33-8773-01d598b83ca7","order_by":7,"name":"Sam van Beljouw","email":"","orcid":"","institution":"Department of Bionanoscience, Delft University of Technology, 2629 HZ Delft, Netherlands","correspondingAuthor":false,"prefix":"","firstName":"Sam","middleName":"van","lastName":"Beljouw","suffix":""},{"id":296889799,"identity":"08cc3e5a-7de2-46ba-8e61-3c51e9286a50","order_by":8,"name":"Marcin Skwark","email":"","orcid":"https://orcid.org/0000-0002-2022-6766","institution":"InstaDeep","correspondingAuthor":false,"prefix":"","firstName":"Marcin","middleName":"","lastName":"Skwark","suffix":""},{"id":296889800,"identity":"cffb5a4a-d81d-4ec9-89dd-a7d4855f45e5","order_by":9,"name":"Andreas Hougaard Laustsen","email":"","orcid":"https://orcid.org/0000-0001-6918-5574","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Andreas","middleName":"Hougaard","lastName":"Laustsen","suffix":""},{"id":296889801,"identity":"a69d546b-4c18-4dab-8b12-8e2ffe2f2e5f","order_by":10,"name":"Anne Ljungars","email":"","orcid":"https://orcid.org/0000-0002-2158-0601","institution":"Department of Biotechnology and Biomedicine, Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Anne","middleName":"","lastName":"Ljungars","suffix":""},{"id":296889802,"identity":"a833be91-549b-4f98-887c-3c7ff82ab058","order_by":11,"name":"Erwin Schoof","email":"","orcid":"https://orcid.org/0000-0002-3117-7832","institution":"Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Erwin","middleName":"","lastName":"Schoof","suffix":""},{"id":296889803,"identity":"1a8972f6-45ee-46b1-81ce-bfc851431573","order_by":12,"name":"Jeroen Van Goey","email":"","orcid":"","institution":"InstaDeep","correspondingAuthor":false,"prefix":"","firstName":"Jeroen","middleName":"Van","lastName":"Goey","suffix":""},{"id":296889804,"identity":"716e4eb6-2d6f-4992-ab82-f8ccd2e69170","order_by":13,"name":"Ulrich auf dem Keller","email":"","orcid":"","institution":"Department of Biotechnology and Biomedicine, Technical University of Denmark","correspondingAuthor":false,"prefix":"","firstName":"Ulrich","middleName":"auf dem","lastName":"Keller","suffix":""},{"id":296889805,"identity":"92113c4c-6202-4277-ac6a-5f87955ce032","order_by":14,"name":"Karim Beguir","email":"","orcid":"","institution":"InstaDeep","correspondingAuthor":false,"prefix":"","firstName":"Karim","middleName":"","lastName":"Beguir","suffix":""},{"id":296889806,"identity":"b2f6547e-3d7f-4c59-ae69-fdd78047f6f8","order_by":15,"name":"Nicolas Lopez Carranza","email":"","orcid":"","institution":"InstaDeep","correspondingAuthor":false,"prefix":"","firstName":"Nicolas","middleName":"Lopez","lastName":"Carranza","suffix":""},{"id":296889807,"identity":"3747c432-5e83-407d-a7e7-65f6d2c5ea76","order_by":16,"name":"Stan Brouns","email":"","orcid":"","institution":"Department of Bionanoscience, Delft University of Technology, 2629 HZ Delft, Netherlands","correspondingAuthor":false,"prefix":"","firstName":"Stan","middleName":"","lastName":"Brouns","suffix":""}],"badges":[],"createdAt":"2023-09-21 13:57:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3376248/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3376248/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s42256-025-01019-5","type":"published","date":"2025-03-31T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":79645925,"identity":"5c2a618e-bede-4fa1-851d-5b7ecdfed785","added_by":"auto","created_at":"2025-04-01 07:05:55","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5323797,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3376248/v1_covered_9c2a3591-e4a3-4fe9-a5c7-c520818002a2.pdf"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nKE, JVG, WW, MS, KB, and NC are employees of InstaDeep, 5 Merchant Square, London, UK. The remaining authors declare no conflicts of interest.","formattedTitle":"De novo peptide sequencing with InstaNovo: Diffusion-powered, peptide identification for large scale proteomics experiments","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"
[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3376248/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3376248/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Bottom-up mass spectrometry-based proteomics is challenged by the task of identifying the peptide that generates a tandem mass spectrum. Traditional methods that rely on known peptide sequence databases are limited and may not be applicable in certain contexts. De novo peptide sequencing, which assigns peptide sequences to the spectra without prior information, is valuable for various biological applica-tions; yet, due to a lack of accuracy, it remains challenging to apply this approach in many situations. Here, we introduce InstaNovo, a transformer neural network with the ability to translate fragment ion peaks into the sequence of amino acids that make up the studied pep-tide(s). The model was trained on 28 million labelled spectra matched to 742k human peptides from the ProteomeTools project. We demon-strate that InstaNovo outperforms current state-of-the-art methods on benchmark datasets and showcase its utility in several applications. Building upon human intuition, we also introduce InstaNovo+, a multi-nomial diffusion model that further improves performance by iterative refinement of predicted sequences. Using these models, we could de novo sequence antibody-based therapeutics with unprecedented cov-erage, discover novel peptides, and detect unreported organisms in different datasets, thereby expanding the scope and detection rate of proteomics searches. Finally, we could experimentally validate tryp-tic and non-tryptic peptides with targeted proteomics, demonstrat-ing the fidelity of our predictions. Our models unlock a plethora of opportunities across different scientific domains, such as direct protein sequencing, immunopeptidomics, and exploration of the dark proteome.","manuscriptTitle":"De novo peptide sequencing with InstaNovo: Diffusion-powered, peptide identification for large scale proteomics experiments","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-02 06:34:05","doi":"10.21203/rs.3.rs-3376248/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"
[email protected]","identity":"nature-machine-intelligence","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"natmachintell","sideBox":"Learn more about [Nature Machine Intelligence](http://www.nature.com/natmachintell/)","snPcode":"","submissionUrl":"","title":"Nature Machine Intelligence","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"f83b4698-7a4b-4e41-af66-963f8acb012d","owner":[],"postedDate":"May 2nd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":31315513,"name":"Biological sciences/Computational biology and bioinformatics/Proteome informatics"},{"id":31315514,"name":"Biological sciences/Computational biology and bioinformatics/Machine learning"}],"tags":[],"updatedAt":"2025-04-01T07:05:40+00:00","versionOfRecord":{"articleIdentity":"rs-3376248","link":"https://doi.org/10.1038/s42256-025-01019-5","journal":{"identity":"nature-machine-intelligence","isVorOnly":false,"title":"Nature Machine Intelligence"},"publishedOn":"2025-03-31 04:00:00","publishedOnDateReadable":"March 31st, 2025"},"versionCreatedAt":"2024-05-02 06:34:05","video":"","vorDoi":"10.1038/s42256-025-01019-5","vorDoiUrl":"https://doi.org/10.1038/s42256-025-01019-5","workflowStages":[]},"version":"v1","identity":"rs-3376248","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3376248","identity":"rs-3376248","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.