서브메뉴
검색
Computational Metabolomics for Bioactivity-Driven Classification in Natural Products Drug Discovery and Opioid Detection
Computational Metabolomics for Bioactivity-Driven Classification in Natural Products Drug Discovery and Opioid Detection
Detailed Information
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202103613
- ISBN
- 9798315733874
- DDC
- 615
- 서명/저자
- Computational Metabolomics for Bioactivity-Driven Classification in Natural Products Drug Discovery and Opioid Detection
- 발행사항
- [Sl] : The University of Wisconsin - Madison, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 175 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-11, Section: B.
- 주기사항
- Advisor: Bugni, Tim S.
- 학위논문주기
- Thesis (Ph.D.)--The University of Wisconsin - Madison, 2025.
- 초록/해제
- 요약The rise of multidrug-resistant (MDR) pathogens and the proliferation of synthetic opioids pose urgent threats to global health, demanding innovative approaches in both drug discovery and forensic analysis. This dissertation presents a multidisciplinary framework that leverages high-resolution mass spectrometry, yeast chemical genomics (YCG), and machine learning (ML) to address these dual challenges across two different but connected domains: natural product discovery and clinical toxicology (Chapter 1).To accelerate the identification of structurally novel and mechanistically distinct antifungal agents, we developed a high-throughput screening pipeline that integrates LC-MS/MS-based metabolomics with YCG profiling (Chapter 2). Nearly 40,000 bacterial extract fractions from diverse microbiomes were screened for antifungal activity, with hits prioritized based on both chemical-genetic interaction signatures and computational metabolomics identifications. This dual-platform approach enabled functional dereplication by linking mass spectral features to both known drug classes and unexplored bioactive compounds, streamlining the prioritization of lead scaffolds for further development.To further enhance the dereplication and prioritization of small molecules, we trained ML classifiers on in-silico MS/MS fingerprints to assign natural products to 21 pharmacophore-defined drug classes (Chapter 3). These classifiers achieved 93% accuracy in multiclass tasks and consistently outperformed state-of-the-art tools (e.g., CANOPUS) across in-silico, GNPS, and experimental datasets. When applied to microbial extracts, the models enabled rapid classification of bioactivity, even in the absence of direct structural matches, expanding the accessible chemical and functional space for antimicrobial discovery.The second major focus of this dissertation is the development of ML-based models for opioid detection from clinical LC-MS/MS data (Chapter 4). Classifiers for morphinan, fentanyl, and nitazene opioids were trained on simulated molecular fingerprints and validated using GNPS spectra, CDC FAS Kit reference materials, and anonymized clinical blood and urine samples. Morphinan and fentanyl models demonstrated exceptional performance, increasing compound detection and sample coverage by 90-600% and 33-400%, respectively, compared to traditional spectral library searches. Although the nitazene classifier performed well in controlled settings, clinical application highlighted the need for improved coverage and training data.Collectively, these studies illustrate how integrating high-resolution metabolomics with machine learning and functional genomics can transform both small molecule discovery and public health surveillance, enabling scalable, structure-informed detection of bioactive natural products and emerging synthetic opioids.
- 일반주제명
- Pharmaceutical sciences
- 일반주제명
- Chemistry
- 일반주제명
- Computational chemistry
- 일반주제명
- Bioinformatics
- 일반주제명
- Analytical chemistry
- 키워드
- Drug discovery
- 키워드
- Machine learning
- 키워드
- Metabolomics
- 키워드
- Natural products
- 기타저자
- The University of Wisconsin - Madison Pharmaceutical Sciences
- 기본자료저록
- Dissertations Abstracts International. 86-11B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017357886
■00520260202103613
■006m o d
■007cr#unu||||||||
■020 ▼a9798315733874
■035 ▼a(MiAaPQ)AAI32043908
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a615
■1001 ▼aBrittin, Nathaniel Jacob.
■24510▼aComputational Metabolomics for Bioactivity-Driven Classification in Natural Products Drug Discovery and Opioid Detection
■260 ▼a[Sl]▼bThe University of Wisconsin - Madison▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a175 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-11, Section: B.
■500 ▼aAdvisor: Bugni, Tim S.
■5021 ▼aThesis (Ph.D.)--The University of Wisconsin - Madison, 2025.
■520 ▼aThe rise of multidrug-resistant (MDR) pathogens and the proliferation of synthetic opioids pose urgent threats to global health, demanding innovative approaches in both drug discovery and forensic analysis. This dissertation presents a multidisciplinary framework that leverages high-resolution mass spectrometry, yeast chemical genomics (YCG), and machine learning (ML) to address these dual challenges across two different but connected domains: natural product discovery and clinical toxicology (Chapter 1).To accelerate the identification of structurally novel and mechanistically distinct antifungal agents, we developed a high-throughput screening pipeline that integrates LC-MS/MS-based metabolomics with YCG profiling (Chapter 2). Nearly 40,000 bacterial extract fractions from diverse microbiomes were screened for antifungal activity, with hits prioritized based on both chemical-genetic interaction signatures and computational metabolomics identifications. This dual-platform approach enabled functional dereplication by linking mass spectral features to both known drug classes and unexplored bioactive compounds, streamlining the prioritization of lead scaffolds for further development.To further enhance the dereplication and prioritization of small molecules, we trained ML classifiers on in-silico MS/MS fingerprints to assign natural products to 21 pharmacophore-defined drug classes (Chapter 3). These classifiers achieved 93% accuracy in multiclass tasks and consistently outperformed state-of-the-art tools (e.g., CANOPUS) across in-silico, GNPS, and experimental datasets. When applied to microbial extracts, the models enabled rapid classification of bioactivity, even in the absence of direct structural matches, expanding the accessible chemical and functional space for antimicrobial discovery.The second major focus of this dissertation is the development of ML-based models for opioid detection from clinical LC-MS/MS data (Chapter 4). Classifiers for morphinan, fentanyl, and nitazene opioids were trained on simulated molecular fingerprints and validated using GNPS spectra, CDC FAS Kit reference materials, and anonymized clinical blood and urine samples. Morphinan and fentanyl models demonstrated exceptional performance, increasing compound detection and sample coverage by 90-600% and 33-400%, respectively, compared to traditional spectral library searches. Although the nitazene classifier performed well in controlled settings, clinical application highlighted the need for improved coverage and training data.Collectively, these studies illustrate how integrating high-resolution metabolomics with machine learning and functional genomics can transform both small molecule discovery and public health surveillance, enabling scalable, structure-informed detection of bioactive natural products and emerging synthetic opioids.
■590 ▼aSchool code: 0262.
■650 4▼aPharmaceutical sciences
■650 4▼aChemistry
■650 4▼aComputational chemistry
■650 4▼aBioinformatics
■650 4▼aAnalytical chemistry
■653 ▼aChemical genomics
■653 ▼aDrug discovery
■653 ▼aMachine learning
■653 ▼aMass spectrometry
■653 ▼aMetabolomics
■653 ▼aNatural products
■690 ▼a0572
■690 ▼a0485
■690 ▼a0219
■690 ▼a0486
■690 ▼a0715
■71020▼aThe University of Wisconsin - Madison▼bPharmaceutical Sciences.
■7730 ▼tDissertations Abstracts International▼g86-11B.
■790 ▼a0262
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17357886▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
Preview
Export
ChatGPT Discussion
AI Recommended Related Books
Подробнее информация.
- Бронирование
- не существует
- моя папка
- Первый запрос зрения
- Non-Book Loan Application
- Nighttime Book Loan Application
Available after logging in.


