Item

Abstract

Over the last year the research on epigenetics has been intensified, as the involvement of epigenetic mechanisms in human diseases has become evident. With the generation of the first genome-wide datasets, the demand for appropriate tools for their automated analysis begins to form. During the same period an exponential growth of public available genome assemblies, has been triggered by the development and spread of efficient DNA sequencing technologies. In consequence of this situation a new scientific discipline is forming, for which the term �comparative epigenomics� can be coined. In this thesis a novel computational pipeline is presented, which enables the in silico analysis of the evolutionary development of large scale datasets of genomic regions. In a pilot study on datasets of CpG islands on human chromosome 21, this pipeline is used to study the conservation of the DNA sequence, DNA structure, DNA base composition and of putative epigenetic footprints of DNA methylation in homologous regions in genome assemblies of 17 vertebrate species. Furthermore, recent machine learning approaches are applied to investigate in how far computational methods for the prediction of the methylation status of CpG islands can be enhanced, by information from the comparative epigenomics field.