서브메뉴
검색
Leveraging Large Scale Electronic Health Record Databases to Improve Biological Analysis
Leveraging Large Scale Electronic Health Record Databases to Improve Biological Analysis
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202104739
- ISBN
- 9798290651958
- DDC
- 610.285
- 서명/저자
- Leveraging Large Scale Electronic Health Record Databases to Improve Biological Analysis
- 발행사항
- [Sl] : Stanford University, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 144 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
- 주기사항
- Advisor: Aghaeepour, Nima.
- 학위논문주기
- Thesis (Ph.D.)--Stanford University, 2025.
- 초록/해제
- 요약This dissertation explores innovative approaches to leverage electronic health record (EHR) databases for enhanced biological analysis. Through several interconnected studies, I demonstrate how integrating clinical and biological data can overcome the limitations inherent to each data type when analyzed in isolation. First, I introduce COMET (Clinical and Omics Multi-Modal Analysis Enhanced with Transfer Learning), a novel deep learning framework that utilizes transfer learning from large EHR databases to improve the analysis of high-dimensional omics data from smaller cohorts. I demonstrate COMET's application to pregnancy outcomes and cancer prognosis, showing how it consistently outperforms traditional analytical approaches. We demonstrate both improved predictive modeling and improved biological discovery through use of the COMET framework. I then demonstrate COMET's utility to diabetic retinopathy, validating known disease mechanisms while identifying novel potential biomarkers and therapeutic targets. Lastly, I evaluate different pre-training strategies for EHR foundation models, providing guidance for further development. This work establishes a methodological foundation for integrating clinical and biological data that combines the statistical robustness of large-scale EHR studies with the mechanistic insights of molecular profiling, ultimately accelerating our path toward more personalized, mechanism-based biomedical discoveries.
- 일반주제명
- Medical prognosis
- 일반주제명
- Visualization
- 일반주제명
- Mortality
- 일반주제명
- Bioinformatics
- 기타저자
- Stanford University.
- 기본자료저록
- Dissertations Abstracts International. 87-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017358700
■00520260202104739
■006m o d
■007cr#unu||||||||
■020 ▼a9798290651958
■035 ▼a(MiAaPQ)AAI32149687
■035 ▼a(MiAaPQ)Stanfordmk929xk0922
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a610.285
■1001 ▼aMataraso, Samson Joel.
■24510▼aLeveraging Large Scale Electronic Health Record Databases to Improve Biological Analysis
■260 ▼a[Sl]▼bStanford University▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a144 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-03, Section: B.
■500 ▼aAdvisor: Aghaeepour, Nima.
■5021 ▼aThesis (Ph.D.)--Stanford University, 2025.
■520 ▼aThis dissertation explores innovative approaches to leverage electronic health record (EHR) databases for enhanced biological analysis. Through several interconnected studies, I demonstrate how integrating clinical and biological data can overcome the limitations inherent to each data type when analyzed in isolation. First, I introduce COMET (Clinical and Omics Multi-Modal Analysis Enhanced with Transfer Learning), a novel deep learning framework that utilizes transfer learning from large EHR databases to improve the analysis of high-dimensional omics data from smaller cohorts. I demonstrate COMET's application to pregnancy outcomes and cancer prognosis, showing how it consistently outperforms traditional analytical approaches. We demonstrate both improved predictive modeling and improved biological discovery through use of the COMET framework. I then demonstrate COMET's utility to diabetic retinopathy, validating known disease mechanisms while identifying novel potential biomarkers and therapeutic targets. Lastly, I evaluate different pre-training strategies for EHR foundation models, providing guidance for further development. This work establishes a methodological foundation for integrating clinical and biological data that combines the statistical robustness of large-scale EHR studies with the mechanistic insights of molecular profiling, ultimately accelerating our path toward more personalized, mechanism-based biomedical discoveries.
■590 ▼aSchool code: 0212.
■650 4▼aElectronic health records
■650 4▼aMedical prognosis
■650 4▼aVisualization
■650 4▼aMortality
■650 4▼aBioinformatics
■653 ▼aElectronic health record
■690 ▼a0715
■71020▼aStanford University.
■7730 ▼tDissertations Abstracts International▼g87-03B.
■790 ▼a0212
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17358700▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
