서브메뉴
검색
Developing Differentiable Toolkits for Computational Biology
Developing Differentiable Toolkits for Computational Biology
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260311091529.5
- ISBN
- 9798280710337
- DDC
- 001
- 저자명
- Ho Pahng, Seong
- 서명/저자
- Developing Differentiable Toolkits for Computational Biology / Seong Ho Pahng
- 발행사항
- [Sl] : Harvard University, 2025
- 형태사항
- 1 electronic resource (128 pages)
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-12, Section: B.
- 주기사항
- Advisors: Hormoz, Sahand Committee members: Shakhnovich, Eugene; Mason, Jarad.
- 학위논문주기
- - Ph.D. : Harvard University, 2025.
- 초록/해제
- 요약In any biological experiment, no matter how sophisticated, we capture only a small, noisy glimpse of a complex underlying process. For dry-lab researchers, interpreting these messy data raises a fundamental question: should one strive for a mechanistic understanding of the biological processes involved, or simply focus on data analysis for a specific task? This dissertation presents three computational tools that emerged from our attempt to strike a balance between these two extremes when modeling novel data.Chapter 1 presents two stochastic models that address the major contamination issues in probebased bacterial single-cell sequencing: spurious unique molecular identifier (UMI) counts and the difficulty of distinguishing genuine cellular signals from noise. By modeling two specific steps of the 10x sequencing pipeline, these methods accurately infer true UMI counts and identify real cells, enabling downstream single-cell analyses that revealed heterogeneous toxin expression in isogenic C. perfringens populations.Chapter 2 employs a deep learning technique to predict cellular responses in Perturb-seq experiments. We posit that the intermediate biological adaptations governing these responses are driven by gene regulatory networks composed of directed, nonreciprocal interactions. To model such interactions, we propose a novel directed graph neural network (CoED) along with a new Laplacian (Fuzzy graph Laplacian) that better captures directional effects. We show that learning both the edge directions and the CoED parameters simultaneously improves predictive performance over existing methods.Chapter 3 presents a differentiable in silico morphogenesis framework that learns to transform a spherical arrangement of point clouds into any desired 3D shape. To compare 3D objects in a manner invariant to index permutations, density, and orientation, we design a loss function that operates in the spectral domain. We also propose a neural network-based force model in which individual agents learn to interact so that, collectively, the system forms the target shape.
- 언어주기
- English
- 일반주제명
- Biology
- 일반주제명
- Applied mathematics
- 일반주제명
- Bioinformatics
- 키워드
- Data analysis
- 기타저자
- Harvard University Chemistry and Chemical Biology
- 기본자료저록
- Dissertations Abstracts International. 86-12B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260311s2025 us eng d■001000017357553
■00520260311091529.5
■006m o d
■007cr|nu||||||||
■020 ▼a9798280710337
■040 ▼aMiAaPQD▼beng▼cMiAaPQD▼erda
■082 ▼a001
■1001 ▼aHo Pahng, Seong ▼eauthor.▼0(orcid)0000-0002-9247-7425
■24510▼aDeveloping Differentiable Toolkits for Computational Biology ▼cSeong Ho Pahng
■260 ▼a[Sl]▼bHarvard University▼c2025
■264 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a1 electronic resource (128 pages)
■336 ▼atext▼btxt▼2rdacontent
■337 ▼acomputer▼bc▼2rdamedia
■338 ▼aonline resource▼bcr▼2rdacarrier
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-12, Section: B.
■500 ▼aAdvisors: Hormoz, Sahand Committee members: Shakhnovich, Eugene; Mason, Jarad.
■5021 ▼bPh.D.▼cHarvard University▼d2025.
■520 ▼aIn any biological experiment, no matter how sophisticated, we capture only a small, noisy glimpse of a complex underlying process. For dry-lab researchers, interpreting these messy data raises a fundamental question: should one strive for a mechanistic understanding of the biological processes involved, or simply focus on data analysis for a specific task? This dissertation presents three computational tools that emerged from our attempt to strike a balance between these two extremes when modeling novel data.Chapter 1 presents two stochastic models that address the major contamination issues in probebased bacterial single-cell sequencing: spurious unique molecular identifier (UMI) counts and the difficulty of distinguishing genuine cellular signals from noise. By modeling two specific steps of the 10x sequencing pipeline, these methods accurately infer true UMI counts and identify real cells, enabling downstream single-cell analyses that revealed heterogeneous toxin expression in isogenic C. perfringens populations.Chapter 2 employs a deep learning technique to predict cellular responses in Perturb-seq experiments. We posit that the intermediate biological adaptations governing these responses are driven by gene regulatory networks composed of directed, nonreciprocal interactions. To model such interactions, we propose a novel directed graph neural network (CoED) along with a new Laplacian (Fuzzy graph Laplacian) that better captures directional effects. We show that learning both the edge directions and the CoED parameters simultaneously improves predictive performance over existing methods.Chapter 3 presents a differentiable in silico morphogenesis framework that learns to transform a spherical arrangement of point clouds into any desired 3D shape. To compare 3D objects in a manner invariant to index permutations, density, and orientation, we design a loss function that operates in the spectral domain. We also propose a neural network-based force model in which individual agents learn to interact so that, collectively, the system forms the target shape.
■546 ▼aEnglish
■590 ▼aSchool code: 0084
■650 4▼aBiology
■650 4▼aApplied mathematics
■650 4▼aBioinformatics
■653 ▼aComputational tools
■653 ▼aUnique molecular identifier
■653 ▼aSingle-cell sequencing
■653 ▼aData analysis
■7102 ▼aHarvard University▼bChemistry and Chemical Biology.▼edegree granting institution.
■7201 ▼aHormoz, Sahand▼edegree supervisor.
■7730 ▼tDissertations Abstracts International▼g86-12B.
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17357553▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.


