Statistical Methods for High-throughput Epigenomics Data Restricted; Files Only

Abstract

DNA methylation is an important epigenetic modification that has essential roles in biological and clinical processes including gene regulation, development and disease. Aberrant and unique DNA methylation patterns have been identified in various diseases such as cancer, making DNA methylation an ideal biomarker. Recently, various high-throughput technologies have emerged to measure genome-wide epigenomics profiles. However, due to the novelty of the technologies and special characteristics of the high-throughput DNA methylation data, there lacks rigorous and effective statistical methods to examine DNA methylation thoroughly.

The coherent theme of this dissertation is to develop novel statistical models and data analysis strategies for high-throughput epigenomics data. In particular, I propose several model-based methods for studying DNA methylation and its relationship to disease. My first research topic aims at classifying tumors into different subtypes based on their methylation profiles, which can facilitate the application of precision medicine on patients. In practice, the data obtained from clinical samples are mixed signals. The proportion of cancer cells in the mixture, known as the tumor purity, will bias the clustering results if not properly accounted for. In this work, I develop a model-based clustering method to infer tumor subtypes with the consideration of tumor purity.

Moving from solid tumor samples to blood assay, my second research topic aims at using cell-free DNA (cfDNA) methylation data to detect disease. Recent researches start to exploit the epigenetic information on cfDNA, which could have broad applications. In this work, I provide thorough reviews and discussions on the statistical method developments and data analysis strategies for using cfDNA epigenetic profiles, in particular DNA methylation, to construct disease diagnostic models.

Along the trajectory of studying cfDNA, my third research topic aims at investigating another type of epigenetic marker: 5-hydroxymethylcytosine (5hmC). Currently, little is known about the 5hmC epigenetic profile on cfDNA. Here, I investigate the genome-wide alteration of cfDNA 5hmC in young healthy subjects, old healthy subjects and late onset Alzheimer’s disease (AD) patients. This is the first investigation, both experimentally and computationally, to study the cfDNA 5hmC profile of neurodegenerative disease and its potential as a diagnostic biomarker.

Table of Contents

1 Introduction 1

1.1 DNA Methylation............................. 2

1.2 Tumor Subtype Classification ...................... 3

1.3 Cell-FreeDNA .............................. 4

1.4 Overarching Goal and Outline...................... 6 

2 Accounting for Tumor Purity Improves Cancer Subtype Classification from DNA Methylation Data 9

2.1 Introduction................................ 10

2.2 Materials and Methods.......................... 12

2.3 Results................................... 17

2.4 Discussion................................. 28 

3 Disease Prediction by Cell-Free DNA Methylation 30

3.1 Introduction................................ 31

3.2 The Cause of Alteration of cfDNA Methylation in Disease . . . . . . 32 

3.3 ExistingWorks .............................. 33

3.4 Methods.................................. 34 

3.5 Real data results ............................. 52

3.6 Discussion and Conclusion........................ 56

3.7 Key Points................................. 58

3.8 Methods Availability ........................... 59 

4 Cell-Free 5hmC in Alzheimer’s Disease Patients 60

4.1 Introduction................................ 61

4.2 Materials and Methods.......................... 63 

4.3 Results................................... 67 

4.4 Discussion................................. 74

5 Conclusion and Future Research Plan 79

5.1 Conclusion................................. 80

5.2 Future Research Plan........................... 80 

About this Dissertation

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Laney Graduate School
Department	Biostatistics
Degree	Ph.D.
Submission	Dissertation
Language	English
Research Field	Biology, Bioinformatics Statistics Biology, Biostatistics
Parola chiave	Epigenetics Microarray Bioinformatics Next Generation Sequencing Cell-free DNA Computational Biology DNA Methylation
Committee Chair / Thesis Advisor	Hao Wu, Emory University
Committee Members	Peng Jin, Emory University Zhaohui Steve Qin, Emory University Karen Conneely, Emory University

Ultima modifica

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	File download under embargo until 20 August 2025	2019-06-10 10:29:21 -0400	File download under embargo until 20 August 2025

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions