{"id":{"repo_id":"msu","oai_identifier":"oai:d.lib.msu.edu:etd_50755"},"canonical_url":"https://search.ndltd.org/etd/msu/oai:d.lib.msu.edu:etd_50755","repository":{"repo_id":"msu","name":"Michigan State University","base_url":"https://d.lib.msu.edu/oai"},"display":{"title":"AI accelerated collisional cross section prediction for high throughput metabolite identification","abstract":"Metabolomics refers to the collective characterization of small organic molecules in a biological sample. While instrumentation and software continues to improve for metabolomics studies, the fraction of annotated signals in untargeted metabolomics experiments remain small. Translating features to metabolite identities present a major bottleneck, confounded by the lack of authentic standards to build comprehensive experimental databases. I illustrate the development of collisional cross section (CCS) prediction methods through deduction from theory and induction from available data. The theoretical CCS prediction involves multistep modeling of conformational ensemble followed by simulation of ion mobility. The advanced computational chemistry operations were automated using the AutoGraph conformational clustering protocol and implementation of the workflow in Snakemake. In a complementary approach, I applied a graph convolutional deep Bayesian neural net to predict CCS values and their uncertainty values. The quantified uncertainty was used to guide ab initio prediction of CCS values in an active learning strategy. The developed methodologies lay the foundation to a continuously refining in silico CCS library to aid in metabolite annotation.","abstract_html":"Metabolomics refers to the collective characterization of small organic molecules in a biological sample. While instrumentation and software continues to improve for metabolomics studies, the fraction of annotated signals in untargeted metabolomics experiments remain small. Translating features to metabolite identities present a major bottleneck, confounded by the lack of authentic standards to build comprehensive experimental databases. I illustrate the development of collisional cross section (CCS) prediction methods through deduction from theory and induction from available data. The theoretical CCS prediction involves multistep modeling of conformational ensemble followed by simulation of ion mobility. The advanced computational chemistry operations were automated using the AutoGraph conformational clustering protocol and implementation of the workflow in Snakemake. In a complementary approach, I applied a graph convolutional deep Bayesian neural net to predict CCS values and their uncertainty values. The quantified uncertainty was used to guide ab initio prediction of CCS values in an active learning strategy. The developed methodologies lay the foundation to a continuously refining in silico CCS library to aid in metabolite annotation.","abstract_has_math":false,"creators":["Tanemura, Kiyoto Aramis"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Merz, Kenneth M.","Wilson, Angela K.","Hong, Heedeok","Wei, Guowei"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2022,"date_issued":"2022","date_published":"2022","updated_at":"2026-11-18T18:57:07Z","subjects":["Chemistry","Computational chemistry","Metabolites","Ion mobility mass spectrometry","Machine learning"],"languages":["English"],"rights":["Attribution-ShareAlike 4.0 International"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:50755","isbn:9798351457123","oclc:on1373601493","oclc:1373601493","umi:29394163","local:Tanemura_grad.msu_0128D_19439"],"render_values":[{"text":"etd:50755","href":null,"code":true},{"text":"isbn:9798351457123","href":null,"code":true},{"text":"oclc:on1373601493","href":null,"code":true},{"text":"oclc:1373601493","href":null,"code":true},{"text":"umi:29394163","href":null,"code":true},{"text":"local:Tanemura_grad.msu_0128D_19439","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/doi:10.25335/q5nn-0z16","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/doi:10.25335/q5nn-0z16","record_label":"Repository record","record_kind":"doi","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_50755","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Merz, Kenneth M.","Wilson, Angela K.","Hong, Heedeok","Wei, Guowei"]},{"key":"dc:creator","label":"Author","values":["Tanemura, Kiyoto Aramis"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2022"]},{"key":"dc:relation","label":"Dc Relation","values":["Electronic Theses & Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["Text","Theses"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Chemistry","Computational chemistry","Metabolites","Ion mobility mass spectrometry","Machine learning"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["English"]},{"key":"dc:rights","label":"Dc Rights","values":["Attribution-ShareAlike 4.0 International"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:50755","isbn:9798351457123","oclc:on1373601493","oclc:1373601493","umi:29394163","local:Tanemura_grad.msu_0128D_19439","https://doi.org/doi:10.25335/q5nn-0z16"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Metabolomics refers to the collective characterization of small organic molecules in a biological sample. While instrumentation and software continues to improve for metabolomics studies, the fraction of annotated signals in untargeted metabolomics experiments remain small. Translating features to metabolite identities present a major bottleneck, confounded by the lack of authentic standards to build comprehensive experimental databases. I illustrate the development of collisional cross section (CCS) prediction methods through deduction from theory and induction from available data. The theoretical CCS prediction involves multistep modeling of conformational ensemble followed by simulation of ion mobility. The advanced computational chemistry operations were automated using the AutoGraph conformational clustering protocol and implementation of the workflow in Snakemake. In a complementary approach, I applied a graph convolutional deep Bayesian neural net to predict CCS values and their uncertainty values. The quantified uncertainty was used to guide ab initio prediction of CCS values in an active learning strategy. The developed methodologies lay the foundation to a continuously refining in silico CCS library to aid in metabolite annotation.","Thesis (Ph. D.)--Michigan State University. Chemistry, 2022","Includes bibliographical references (pages 77-104)"]},{"key":"dc:format","label":"Dc Format","values":["vi, 104 pages","application/pdf"]},{"key":"dc:title","label":"Title","values":["AI accelerated collisional cross section prediction for high throughput metabolite identification"]}]}],"canonical_facts":{"dc:contributor":["Merz, Kenneth M.","Wilson, Angela K.","Hong, Heedeok","Wei, Guowei"],"dc:creator":["Tanemura, Kiyoto Aramis"],"dc:date":["2022"],"dc:description":["Metabolomics refers to the collective characterization of small organic molecules in a biological sample. While instrumentation and software continues to improve for metabolomics studies, the fraction of annotated signals in untargeted metabolomics experiments remain small. Translating features to metabolite identities present a major bottleneck, confounded by the lack of authentic standards to build comprehensive experimental databases. I illustrate the development of collisional cross section (CCS) prediction methods through deduction from theory and induction from available data. The theoretical CCS prediction involves multistep modeling of conformational ensemble followed by simulation of ion mobility. The advanced computational chemistry operations were automated using the AutoGraph conformational clustering protocol and implementation of the workflow in Snakemake. In a complementary approach, I applied a graph convolutional deep Bayesian neural net to predict CCS values and their uncertainty values. The quantified uncertainty was used to guide ab initio prediction of CCS values in an active learning strategy. The developed methodologies lay the foundation to a continuously refining in silico CCS library to aid in metabolite annotation.","Thesis (Ph. D.)--Michigan State University. Chemistry, 2022","Includes bibliographical references (pages 77-104)"],"dc:format":["vi, 104 pages","application/pdf"],"dc:identifier":["etd:50755","isbn:9798351457123","oclc:on1373601493","oclc:1373601493","umi:29394163","local:Tanemura_grad.msu_0128D_19439","https://doi.org/doi:10.25335/q5nn-0z16"],"dc:language":["English"],"dc:relation":["Electronic Theses & Dissertations"],"dc:rights":["Attribution-ShareAlike 4.0 International"],"dc:subject":["Chemistry","Computational chemistry","Metabolites","Ion mobility mass spectrometry","Machine learning"],"dc:title":["AI accelerated collisional cross section prediction for high throughput metabolite identification"],"dc:type":["Text","Theses"]},"updated_at":"2026-11-18T18:57:07Z"}