서브메뉴
검색
Efficient Machine Learning Algorithms for Protein Structure Determination With Single-Particle Cryo-Electron Microscopy
Efficient Machine Learning Algorithms for Protein Structure Determination With Single-Particle Cryo-Electron Microscopy
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202104851
- ISBN
- 9798288816444
- DDC
- 572.86
- 서명/저자
- Efficient Machine Learning Algorithms for Protein Structure Determination With Single-Particle Cryo-Electron Microscopy
- 발행사항
- [Sl] : Stanford University, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 181 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-02, Section: B.
- 주기사항
- Advisor: Dunne, Mike.
- 학위논문주기
- Thesis (Ph.D.)--Stanford University, 2025.
- 초록/해제
- 요약Most biological processes responsible for life are carried out by molecular machines called proteins. Knowledge about their three-dimensional structures is key for understanding their functions and for designing new therapeutic compounds, but turning experimental measurements into structural information constitutes a remarkably challenging computational task. In particular, in single-particle cryo-electron microscopy (cryo-EM), an ensemble of three-dimensional (3D) density fields must be derived from a set of noisy and randomly oriented projections. In this thesis, we develop novel algorithms for reconstructing the 3D structures of proteins and other biomolecules from cryo-EM data. Our work focuses on the problem of jointly estimating structures and viewing angles using neural-based approaches. With cryoAI and cryoFIRE, we demonstrate the benefits of amortized inference in terms of computation time for homogeneous and heterogeneous reconstruction on experimental cryo-EM datasets. We also shed light on the limitations of amortized inference in highly noisy regimes and introduce, in cryoDRGN-AI, a two-stage orientation estimation strategy which we validate on challenging experimental datasets. Finally, this thesis describes an extension of cryoDRGN-AI, Hydra, for handling strong compositional heterogeneity.
- 일반주제명
- Ribonucleic acid--RNA
- 일반주제명
- Electrical engineering
- 기타저자
- Stanford University.
- 기본자료저록
- Dissertations Abstracts International. 87-02B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017359221
■00520260202104851
■006m o d
■007cr#unu||||||||
■020 ▼a9798288816444
■035 ▼a(MiAaPQ)AAI32200964
■035 ▼a(MiAaPQ)Stanfordkq162cq2878
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a572.86
■1001 ▼aLevy, Axel Louis Romain.
■24510▼aEfficient Machine Learning Algorithms for Protein Structure Determination With Single-Particle Cryo-Electron Microscopy
■260 ▼a[Sl]▼bStanford University▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a181 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-02, Section: B.
■500 ▼aAdvisor: Dunne, Mike.
■5021 ▼aThesis (Ph.D.)--Stanford University, 2025.
■520 ▼aMost biological processes responsible for life are carried out by molecular machines called proteins. Knowledge about their three-dimensional structures is key for understanding their functions and for designing new therapeutic compounds, but turning experimental measurements into structural information constitutes a remarkably challenging computational task. In particular, in single-particle cryo-electron microscopy (cryo-EM), an ensemble of three-dimensional (3D) density fields must be derived from a set of noisy and randomly oriented projections. In this thesis, we develop novel algorithms for reconstructing the 3D structures of proteins and other biomolecules from cryo-EM data. Our work focuses on the problem of jointly estimating structures and viewing angles using neural-based approaches. With cryoAI and cryoFIRE, we demonstrate the benefits of amortized inference in terms of computation time for homogeneous and heterogeneous reconstruction on experimental cryo-EM datasets. We also shed light on the limitations of amortized inference in highly noisy regimes and introduce, in cryoDRGN-AI, a two-stage orientation estimation strategy which we validate on challenging experimental datasets. Finally, this thesis describes an extension of cryoDRGN-AI, Hydra, for handling strong compositional heterogeneity.
■590 ▼aSchool code: 0212.
■650 4▼aRibonucleic acid--RNA
■650 4▼aElectrical engineering
■653 ▼aMolecular machines
■653 ▼aCompositional heterogeneity
■653 ▼aSingle-particle cryo-electron microscopy
■690 ▼a0544
■690 ▼a0800
■71020▼aStanford University.
■7730 ▼tDissertations Abstracts International▼g87-02B.
■790 ▼a0212
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17359221▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
