PDF, English Download (29MB) | Terms of use

Abstract

There is a complex mutual interplay between three-dimensional (3D) genome organization and cellular activities in bacteria and eukaryotes. The aim of this thesis is to investigate such structure-function relationships.

A main part of this thesis deals with the study of the three-dimensional genome organization using novel techniques for detecting genome-wide contacts using next-generation sequencing. These so called chromatin conformation capture-based methods, such as 5C and Hi-C, give deep insights into the architecture of the genome inside the nucleus, even on a small scale. We shed light on the question how the vastly increasing Hi-C data can generate new insights about the way the genome is organized in 3D.

To this end, we first present the typical Hi-C data processing workflow to obtain Hi-C contact maps and show potential pitfalls in the interpretation of such contact maps using our own data pipeline and publicly available Hi-C data sets. Subsequently, we focus on approaches to modeling 3D genome organization based on contact maps. In this context, a computational tool was developed which interactively visualizes contact maps alongside complementary genomic data tracks. Inspired by machine learning with the help of probabilistic graphical models, we developed a tool that detects the compartmentalization structure within contact maps on multiple scales. In a further project, we propose and test one possible mechanism for the observed compartmentalization within contact maps of genomes across multiple species: Dynamic formation of loops within domains.

In the context of 3D organization of bacterial chromosomes, we present the first direct evidence for global restructuring by long-range interactions of a DNA binding protein. Using Hi-C and live cell imaging of DNA loci, we show that the DNA binding protein Rok forms insulator-like complexes looping the B. subtilis genome over large distances. This biological mechanism agrees with our model based on dynamic formation of loops affecting domain formation in eukaryotic genomes. We further investigate the spatial segregation of the E. coli chromosome during cell division. In particular, we are interested in the positioning of the chromosomal replication origin region based on its interaction with the protein complex MukBEF. We tackle the problem using a combined approach of stochastic and polymer simulations.

Last but not least, we develop a completely new methodology to analyze single molecule localization microscopy images based on topological data analysis. By using this new approach in the analysis of irradiated cells, we are able to show that the topology of repair foci can be categorized depending the distance to heterochromatin.