Item

Abstract

A panel of key transcription factors are the main drivers of the cardiac developmental process and are essential for normal cardiac function. However, limited insights have been generated at a systemslevel about how these factors modulate the overall transcription network, how they act in a combinatorial manner and moreover how they interplay with epigenetic or environmental factors. To tackle these open points, a systems biology approach was chosen such that well-defined highthroughput experiments were used as a starting point and based on their outcome subsequent experiments followed. Gene-focused hypothesis driven laboratory experiments were performed generating time-series data to puzzle down a sequence of transcription factor binding, histone modification and respective gene transcription. Finally, extracted transcription networks were studied and extended based on expression profile disturbances in diseased human hearts. This thesis represents the bioinformatics part of this overall project and aimed to provide the best suitable bioinformatic and statistical data analysis, predicted new transcription networks and proposed consequent laboratory experiments. In the course of the study, five essential datasets were generated and analyzed: (a) genome-wide ChIPchip data of the cardiac transcription factors Srf, Mef2a, Gata4 and Nkx2.5 in cardiac cell culture, (b) microarray gene expression profiles of wildtype and RNAi treated respective cell culture, (c) ChIP-seq data for Srf and histone 3 acetylation in cardiac cell culture, (d) time series ChIP-qPCR data of Srf, p300 and histone 3 acetylation and methylation in mouse hearts at E18.5, P0.5 and P4.5, and (e) gene expression profiles of human diseased hearts. Alongside, a panel of different bioinformatics and statistical methods suitable to analyze these datasets were identified. Their advantages and disadvantages are discussed and knowledge gained in the course of the project is presented. Finally, cardiac transcription networks were predicted based on the wealth of the data which could to a great extent be confirmed in respectively designed follow-up experiments. As a final step, the Cardiac Regulatory INteraction database (CARIN) was built to integrate data from this project with publicly available and relevant datasets as well as cross-species datasets obtained within the European FP6 project “HeartRepair”. The work presented in this thesis was published in two articles (Molecular BioSystems 2008 and PLoS Genetics 2011), one further manuscript is in preparation. All analyses described have been performed in the group of Prof. Dr. Silke R. Sperling at the Max Planck Institute for Molecular Genetics, Berlin.