{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,3]],"date-time":"2026-06-03T23:08:07Z","timestamp":1780528087167,"version":"3.54.1"},"reference-count":50,"publisher":"Oxford University Press (OUP)","issue":"8","license":[{"start":{"date-parts":[[2024,8,17]],"date-time":"2024-08-17T00:00:00Z","timestamp":1723852800000},"content-version":"vor","delay-in-days":16,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["R35GM148219"],"award-info":[{"award-number":["R35GM148219"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,8,2]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Motivation<\/jats:title>\n                  <jats:p>A key challenge in metabolomics is annotating measured spectra from a biological sample with chemical identities. Currently, only a small fraction of measurements can be assigned identities. Two complementary computational approaches have emerged to address the annotation problem: mapping candidate molecules to spectra, and mapping query spectra to molecular candidates. In essence, the candidate molecule with the spectrum that best explains the query spectrum is recommended as the target molecule. Despite candidate ranking being fundamental in both approaches, limited prior works incorporated rank learning tasks in determining the target molecule.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>We propose a novel machine learning model, Ensemble Spectral Prediction (ESP), for metabolite annotation. ESP takes advantage of prior neural network-based annotation models that utilize multilayer perceptron (MLP) networks and Graph Neural Networks (GNNs). Based on the ranking results of the MLP- and GNN-based models, ESP learns a weighting for the outputs of MLP and GNN spectral predictors to generate a spectral prediction for a query molecule. Importantly, training data is stratified by molecular formula to provide candidate sets during model training. Further, baseline MLP and GNN models are enhanced by considering peak dependencies through label mixing and multi-tasking on spectral topic distributions. When trained on the NIST 2020 dataset and evaluated on the relevant candidate sets from PubChem, ESP improves average rank by 23.7% and 37.2% over the MLP and GNN baselines, respectively, demonstrating performance gain over state-of-the-art neural network approaches. However, MLP approaches remain strong contenders when considering top five ranks. Importantly, we show that annotation performance is dependent on the training dataset, the number of molecules in the candidate set and candidate similarity to the target molecule.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Availability and implementation<\/jats:title>\n                  <jats:p>The ESP code, a trained model, and a Jupyter notebook that guide users on using the ESP tool is available at https:\/\/github.com\/HassounLab\/ESP.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btae490","type":"journal-article","created":{"date-parts":[[2024,8,24]],"date-time":"2024-08-24T20:21:31Z","timestamp":1724530891000},"source":"Crossref","is-referenced-by-count":15,"title":["An Ensemble Spectral Prediction (ESP) model for metabolite annotation"],"prefix":"10.1093","volume":"40","author":[{"given":"Xinmeng","family":"Li","sequence":"first","affiliation":[{"name":"Department of Computer Science, Tufts University , Medford, MA, 02155, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yan","family":"Zhou Chen","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Tufts University , Medford, MA, 02155, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Apurva","family":"Kalia","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Tufts University , Medford, MA, 02155, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3386-6076","authenticated-orcid":false,"given":"Hao","family":"Zhu","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Tufts University , Medford, MA, 02155, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Li-ping","family":"Liu","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Tufts University , Medford, MA, 02155, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9477-2199","authenticated-orcid":false,"given":"Soha","family":"Hassoun","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Tufts University , Medford, MA, 02155, United States"},{"name":"Department of Chemical and Biological Engineering, Tufts University , Medford, MA, 02155, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2024,8,17]]},"reference":[{"key":"2024082420210072900_btae490-B1","doi-asserted-by":"crossref","first-page":"W94","DOI":"10.1093\/nar\/gku436","article-title":"Cfm-id: a web server for annotation, spectrum prediction and metabolite identification from tandem mass spectra","volume":"42","author":"Allen","year":"2014","journal-title":"Nucleic Acids Res"},{"key":"2024082420210072900_btae490-B2","doi-asserted-by":"crossref","first-page":"117","DOI":"10.1038\/nmeth0211-117","article-title":"Metabolomics: from small molecules to big ideas","volume":"8","author":"Baker","year":"2011","journal-title":"Nat Methods"},{"key":"2024082420210072900_btae490-B3","first-page":"993","article-title":"Latent dirichlet allocation","volume":"3","author":"Blei","year":"2003","journal-title":"J Mach Learn Res"},{"key":"2024082420210072900_btae490-B4","first-page":"407","author":"Brouard","year":"2017"},{"key":"2024082420210072900_btae490-B5","doi-asserted-by":"crossref","first-page":"W486","DOI":"10.1093\/nar\/gky310","article-title":"Metaboanalyst 4.0: towards more transparent and integrative metabolomics analysis","volume":"46","author":"Chong","year":"2018","journal-title":"Nucleic Acids Res"},{"key":"2024082420210072900_btae490-B6","doi-asserted-by":"crossref","first-page":"12549","DOI":"10.1073\/pnas.1516878112","article-title":"Illuminating the dark matter in metabolomics","volume":"112","author":"da Silva","year":"2015","journal-title":"Proc Natl Acad Sci USA"},{"key":"2024082420210072900_btae490-B7","doi-asserted-by":"crossref","first-page":"1752","DOI":"10.1038\/s41467-023-37446-4","article-title":"Ms2query: reliable and scalable ms2 mass spectra-based analogue search","volume":"14","author":"de Jonge","year":"2023","journal-title":"Nat Commun"},{"key":"2024082420210072900_btae490-B8","doi-asserted-by":"crossref","first-page":"103","DOI":"10.1007\/s11306-022-01963-y","article-title":"Good practices and recommendations for using and benchmarking computational metabolomics metabolite annotation tools","volume":"18","author":"de Jonge","year":"2022","journal-title":"Metabolomics"},{"key":"2024082420210072900_btae490-B9","first-page":"1","volume":"8","author":"D\u00fchrkop","year":"2016"},{"key":"2024082420210072900_btae490-B10","doi-asserted-by":"crossref","first-page":"299","DOI":"10.1038\/s41592-019-0344-8","article-title":"Sirius 4: a rapid tool for turning tandem mass spectra into metabolite structure information","volume":"16","author":"D\u00fchrkop","year":"2019","journal-title":"Nat Methods"},{"key":"2024082420210072900_btae490-B11","doi-asserted-by":"crossref","first-page":"462","DOI":"10.1038\/s41587-020-0740-8","article-title":"Systematic classification of unknown metabolites using high-resolution fragmentation mass spectra","volume":"39","author":"D\u00fchrkop","year":"2021","journal-title":"Nat Biotechnol"},{"key":"2024082420210072900_btae490-B12","doi-asserted-by":"crossref","first-page":"12580","DOI":"10.1073\/pnas.1509788112","article-title":"Searching molecular structure databases with tandem mass spectra using csi: Fingerid","volume":"112","author":"D\u00fchrkop","year":"2015","journal-title":"Proc Natl Acad Sci USA"},{"key":"2024082420210072900_btae490-B13","doi-asserted-by":"crossref","first-page":"965","DOI":"10.1038\/s42256-023-00708-3","article-title":"Annotating metabolite mass spectra with domain-inspired chemical formula transformers","volume":"5","author":"Goldman","year":"2023","journal-title":"Nat Mach Intell"},{"key":"2024082420210072900_btae490-B14","doi-asserted-by":"crossref","first-page":"160","DOI":"10.3390\/metabo10040160","article-title":"Biological filtering and substrate promiscuity prediction for annotating untargeted metabolomics","volume":"10","author":"Hassanpour","year":"2020","journal-title":"Metabolites"},{"key":"2024082420210072900_btae490-B15","doi-asserted-by":"crossref","first-page":"703","DOI":"10.1002\/jms.1777","article-title":"Massbank: a public repository for sharing mass spectral data for life sciences","volume":"45","author":"Horai","year":"2010","journal-title":"J Mass Spectrom"},{"key":"2024082420210072900_btae490-B16","author":"Hu","year":"2019"},{"key":"2024082420210072900_btae490-B17","doi-asserted-by":"crossref","first-page":"e1008724","DOI":"10.1371\/journal.pcbi.1008724","article-title":"Spec2vec: improved mass spectral similarity scoring through learning of structural relationships","volume":"17","author":"Huber","year":"2021","journal-title":"PLoS Comput Biol"},{"key":"2024082420210072900_btae490-B18","doi-asserted-by":"crossref","first-page":"84","DOI":"10.1186\/s13321-021-00558-4","article-title":"Ms2deepscore: a novel deep learning similarity measure to compare tandem mass spectra","volume":"13","author":"Huber","year":"2021","journal-title":"J Cheminform"},{"key":"2024082420210072900_btae490-B19","doi-asserted-by":"crossref","first-page":"221","DOI":"10.1002\/mas.21548","article-title":"Metabolomics toward personalized medicine","volume":"38","author":"Jacob","year":"2019","journal-title":"Mass Spectrom Rev"},{"key":"2024082420210072900_btae490-B20","first-page":"381","author":"Ji","year":"2008"},{"key":"2024082420210072900_btae490-B21","doi-asserted-by":"crossref","first-page":"451","DOI":"10.1038\/nrm.2016.25","article-title":"Metabolomics: beyond biomarkers and towards mechanisms","volume":"17","author":"Johnson","year":"2016","journal-title":"Nat Rev Mol Cell Biol"},{"key":"2024082420210072900_btae490-B22","doi-asserted-by":"crossref","first-page":"D1202","DOI":"10.1093\/nar\/gkv951","article-title":"Pubchem substance and compound databases","volume":"44","author":"Kim","year":"2016","journal-title":"Nucleic Acids Res"},{"key":"2024082420210072900_btae490-B23","doi-asserted-by":"crossref","first-page":"206","DOI":"10.1038\/nature01254","article-title":"Computational systems biology","volume":"420","author":"Kitano","year":"2002","journal-title":"Nature"},{"key":"2024082420210072900_btae490-B24","doi-asserted-by":"crossref","DOI":"10.1101\/2021.08.03.454944","article-title":"Bi-modal variational autoencoders for metabolite identification using tandem mass spectrometry","author":"Kutuzova","year":"2021"},{"key":"2024082420210072900_btae490-B25","author":"Kutuzova","year":"2021"},{"key":"2024082420210072900_btae490-B26","first-page":"2024","author":"Lei","year":"2017"},{"key":"2024082420210072900_btae490-B27","doi-asserted-by":"crossref","first-page":"243","DOI":"10.3390\/metabo10060243","article-title":"Machine learning applications for mass spectrometry-based metabolomics","volume":"10","author":"Liebal","year":"2020","journal-title":"Metabolites"},{"key":"2024082420210072900_btae490-B28","doi-asserted-by":"crossref","first-page":"2077","DOI":"10.1021\/acs.jcim.7b00166","article-title":"Profile-qsar 2.0: kinase virtual screening accuracy comparable to four-concentration ic50s for realistically novel compounds","volume":"57","author":"Martin","year":"2017","journal-title":"J Chem Inf Model"},{"key":"2024082420210072900_btae490-B29","doi-asserted-by":"crossref","first-page":"263","DOI":"10.1038\/nrm3314","article-title":"Metabolomics: the apogee of the omics trilogy","volume":"13","author":"Patti","year":"2012","journal-title":"Nat Rev Mol Cell Biol"},{"key":"2024082420210072900_btae490-B30","doi-asserted-by":"crossref","first-page":"11732","DOI":"10.1021\/ac402689t","article-title":"Development of a standard reference material for metabolomics research","volume":"85","author":"Phinney","year":"2013","journal-title":"Anal Chem"},{"key":"2024082420210072900_btae490-B31","doi-asserted-by":"crossref","first-page":"311","DOI":"10.1089\/cmb.2012.0083","article-title":"Finding maximum colorful subtrees in practice","volume":"20","author":"Rauf","year":"2013","journal-title":"J Comput Biol"},{"key":"2024082420210072900_btae490-B32","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1186\/s13321-016-0115-9","article-title":"Metfrag relaunched: incorporating strategies beyond in silico fragmentation","volume":"8","author":"Ruttkies","year":"2016","journal-title":"J Cheminform"},{"key":"2024082420210072900_btae490-B33","first-page":"593","author":"Schlichtkrull","year":"2018"},{"key":"2024082420210072900_btae490-B34","doi-asserted-by":"crossref","first-page":"1516","DOI":"10.1038\/s41467-019-09550-x","article-title":"Metabolic reaction network-based recursive metabolite annotation for untargeted metabolomics","volume":"10","author":"Shen","year":"2019","journal-title":"Nat Commun"},{"key":"2024082420210072900_btae490-B35","doi-asserted-by":"crossref","first-page":"1793","DOI":"10.3390\/biom11121793","article-title":"Massgenie: a transformer-based deep learning method for identifying small molecules from their mass spectra","volume":"11","author":"Shrivastava","year":"2021","journal-title":"Biomolecules"},{"key":"2024082420210072900_btae490-B36","doi-asserted-by":"crossref","first-page":"747","DOI":"10.1097\/01.ftd.0000179845.53213.39","article-title":"Metlin: a metabolite mass spectral database","volume":"27","author":"Smith","year":"2005","journal-title":"Ther Drug Monit"},{"key":"2024082420210072900_btae490-B37","first-page":"1409","article-title":"A statistical method for evaluating systematic relationships","volume":"38","author":"Sokal","year":"1958","journal-title":"Univ Kansas Sci Bull"},{"key":"2024082420210072900_btae490-B38","doi-asserted-by":"crossref","first-page":"865","DOI":"10.1038\/s41592-022-01486-3","article-title":"Msnovelist: de novo structure generation from mass spectra","volume":"19","author":"Stravs","year":"2022","journal-title":"Nat Methods"},{"key":"2024082420210072900_btae490-B39","doi-asserted-by":"crossref","first-page":"13738","DOI":"10.1073\/pnas.1608041113","article-title":"Topic modeling for untargeted substructure exploration in metabolomics","volume":"113","author":"van Der Hooft","year":"2016","journal-title":"Proc Natl Acad Sci USA"},{"key":"2024082420210072900_btae490-B40","doi-asserted-by":"crossref","first-page":"52","DOI":"10.1016\/j.trac.2015.04.002","article-title":"Using fragmentation trees and mass spectral trees for identifying unknown compounds in metabolomics","volume":"69","author":"Vaniya","year":"2015","journal-title":"Trends Analyt Chem"},{"key":"2024082420210072900_btae490-B41","article-title":"Attention is all you need","volume":"30","author":"Vaswani","year":"2017","journal-title":"Advances in Neural Information Processing Systems, Long Beach, CA, USA"},{"key":"2024082420210072900_btae490-B42","volume-title":"International Conference on Machine Learning","author":"Veli\u010dkovi\u0107","year":"2018"},{"key":"2024082420210072900_btae490-B43","first-page":"977","author":"Wallach","year":"2006"},{"key":"2024082420210072900_btae490-B44","doi-asserted-by":"crossref","first-page":"11692","DOI":"10.1021\/acs.analchem.1c01465","article-title":"Cfm-id 4.0: more accurate esi-ms\/ms spectral prediction and compound identification","volume":"93","author":"Wang","year":"2021","journal-title":"Anal Chem"},{"key":"2024082420210072900_btae490-B45","doi-asserted-by":"crossref","first-page":"828","DOI":"10.1038\/nbt.3597","article-title":"Sharing and community curation of mass spectrometry data with global natural products social molecular networking","volume":"34","author":"Wang","year":"2016","journal-title":"Nat Biotechnol"},{"key":"2024082420210072900_btae490-B46","doi-asserted-by":"crossref","first-page":"700","DOI":"10.1021\/acscentsci.9b00085","article-title":"Rapid prediction of electron\u2013ionization mass spectrometry using neural networks","volume":"5","author":"Wei","year":"2019","journal-title":"ACS Cent Sci"},{"key":"2024082420210072900_btae490-B47","doi-asserted-by":"crossref","first-page":"881","DOI":"10.1038\/s41592-023-01850-x","article-title":"Buddy: molecular formula discovery via bottom-up ms\/ms interrogation","volume":"20","author":"Xing","year":"2023","journal-title":"Nat Methods"},{"key":"2024082420210072900_btae490-B48","volume-title":"International Conference on Machine Learning","author":"Xu","year":"2019"},{"key":"2024082420210072900_btae490-B49","author":"Young","year":"2021"},{"key":"2024082420210072900_btae490-B50","author":"Zhu","year":"2020"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/40\/8\/btae490\/58914502\/btae490.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/40\/8\/btae490\/58914502\/btae490.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,8,24]],"date-time":"2024-08-24T20:21:50Z","timestamp":1724530910000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/doi\/10.1093\/bioinformatics\/btae490\/7734912"}},"subtitle":[],"editor":[{"given":"Jonathan","family":"Wren","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2024,8]]},"references-count":50,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2024,8,2]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btae490","relation":{},"ISSN":["1367-4811"],"issn-type":[{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024,8]]},"published":{"date-parts":[[2024,8]]},"article-number":"btae490"}}