Dissecting cellular states and cell state transitions through integrative analysis of epigenetic dynamics

DSpace Repository


URI: http://hdl.handle.net/10900/57790
Dokumentart: Dissertation
Date: 2014
Language: English
Faculty: 7 Mathematisch-Naturwissenschaftliche Fakultät
Department: Informatik
Advisor: Kohlbacher, Oliver (Prof. Dr.)
Day of Oral Examination: 2014-09-26
DDC Classifikation: 004 - Data processing and computer science
Keywords: Bioinformatik , Genom
License: Publishing license including print on demand
Order a printed copy: Print-on-Demand
Show full item record


Understanding how a single genome that is common to all cells in an organism can give rise to many different and highly specialized, cell types has been one of the major questions in biology over the past century and still many aspects remain unanswered. Over the last 15 years, incredible progress has been made in pinpointing the regulatory mechanisms that establish, maintain, and change cellular identities. In particular, the role of histone modifications and DNA methylation in the spatio-temporal control of gene expression and genome organization has been greatly appreciated. These histone and DNA modifications have been shown to be an integral part of epigenetic control mechanisms. They ensure stable silencing of not-required genes and gene regulatory elements as well as maintenance of active genes and gene regulatory elements that are required in a particular cellular context. In addition to the identification and functional characterization of these mechanisms, the sequencing of many complex genomes and the advent of high-throughput sequencing technology has allowed us to precisely chart the location of all modified histones and methylated bases across the entire genome. In contrast to the genome sequence, the epigenome turns out to be highly variable between distinct cell types and reflective of their specific biology. Comprehensive mapping efforts of the epigenome provide a starting point for understanding the epigenetic basis of macroscopic phenotypes such as distinct cell types and states. Integrative analysis of many different types of histone modifications combined with gene expression data across several distinct cell types also revealed that certain histone modifications can be used to annotate and predict the activity of different types of gene regulatory elements such as promoters or enhancers. This thesis takes advantage of these recent advances in the identifiability of gene regulatory elements and first establishes that the integration of epigenetic and transcriptional data on specific cellular states can be used to gain insights into the underlying regulatory logic maintaining and establishing these states. Second, it demonstrates the utility of this approach to generate experimentally testable hypothesis on the molecular mechanisms mediating specific cell state transitions. While great progress has been made on mapping changes in histone modifications and identification of their demarcated gene regulatory elements, less is known about methylation of the DNA. In particular, it is still unclear to what extend and where DNA methylation changes over the course of human development and what the likely functions of these changes are. To fill this gap, we mapped DNA methylation patterns using whole-genome bisulfite sequencing across 30 distinct human cell and tissue types. Surprisingly, we find that only around 20% of all CpG dinucleotides, the main target of DNA methylation in mammals, change their methylation state across normal development. Interestingly, we find that most differentially methylated regions (DMRs) coincide with gene regulatory elements, such as enhancers or transcription factor binding sites, that are relevant for the biology of a particular cell type. Most of these DMRs are constitutively hyper-methylated and only become hypo-methylated in a cell type where the underlying gene regulatory element becomes relevant. This study not only determined the extent to which DNA methylation is dynamic during normal development, but also established DNA methylation dynamics as an excellent marker of active, cell type-specific gene regulatory elements. The first part of this thesis examined primarily differences in the DNA methylation landscape of distantly related stable steady state cell types and states. This experimental design is similar to the strategy chosen by most previous investigations of histone modification changes. However, much less is known about the epigenetic changes at high temporal resolution during cell fate transitions. To shed more light on the epigenetic remodeling events that establish and distinguish closely related cell types, we investigated the epigenetic and transcriptional dynamics during the in vitro differentiation of human embryonic stem cells to three distinct early-derived populations of each embryonic germ layer. This analysis revealed comprehensive epigenetic remodeling, particularly at enhancers and frequently converse dynamics between different germ layer populations. In addition, we identified an unexpected, epigenetic transition where a repressed locus first transiently gains H3K27me3 at an early stage of differentiation before becoming fully activated at a later time point. While these experiments and analyses were very informative from an observational point of view, it still is unclear in many cases what are the key drivers of the cell state transitions and in particular of the epigenetic remodeling events. To shed light on this important problem, we developed a computational analysis strategy that leverages epigenetic information to identify key transcription factors (TFs) that are likely to mediate epigenetic changes and orchestrate particular cell state transitions. We then applied this approach to a 200-day time series of a novel, in vitro differentiation scheme of human embryonic stem cells towards the neural lineage. This analysis revealed a core network of transcription factors around PAX6 and OTX2 that is active during the entire process of in vitro neural differentiation. Furthermore, these core factors show evidence of differentiation stage-specific co-binding with other TFs as well as binding site redistribution across the genome. These results highlight the utility and general applicability of this framework for integrated epigenetic analysis. In particular, this approach provides a TF-centric perspective on the interpretation of epigenetic changes during cell state transitions. In summary, this work has contributed to our understanding of the dynamic regulation of DNA methylation and provided a high-resolution, in-depth investigation of epigenetic dynamics during differentiation of embryonic stem cells to three distinct, differentiated cell populations. Furthermore, we provided a detailed view on the regulatory network activity and architecture orchestrating in vitro neural differentiation. In addition, this work yielded novel tools and perspectives to interpret the rapidly rising number of epigenetic profiles.

This item appears in the following Collection(s)