Development of computational approaches for the analysis of bisulfite next-generation sequencing data
The scientific contribution of this thesis consists of three articles that have been published in Bioinformatics (Oxford Journals) and Nature Methods and the third article being under review at Leukemia (Nature Publishing Group), respectively. The implications of these articles for the field of computational epigenetics and future perspectives of this research area are discussed. The main challenge within the framework of this thesis was the development of a bioinformatics tool for bisulfite sequencing analysis. The article in Bioinformatics presents the bioinformatics tool B-SOLANA for the analysis of DNA methylation data generated by two-base encoding bisulfite sequencing on the SOLiD platform of Life Technologies. Additionally, benchmark analyses revealed that B-SOLANA exhibits a significantly higher sensitivity and specificity compared to other software approaches which were developed at the same time. The review article in Nature Methods summarizes challenges of bisulfite sequencing analysis as they appear on different high-throughput sequencing platforms. Especially primary analyses including the quality control and mapping of raw sequences are discussed. Furthermore, the article debates the effect of sequencing errors and contaminations on inferred DNA methylation levels and recommends the most appropriate way to analyze this type of data. This review is a helpful reference for the analysis of DNA methylation by high-throughput sequencing, a currently rapidly developing research area. The third article, which has been submitted to Leukemia, comprises the analysis of a DNA methylome of the DAUDI cell line at single base resolution. On the genetic level, this endemic Burkitt Lymphoma cell line is characterized by the presence of the hallmark IG-MYC translocation. Recent publications about this cell line suggested a high number of DNA methylation changes. However, until now only array-based studies were published, which have concentrated their focus on loci-specific DNA methylation patterns. We showed that the mechanisms of DNA methylation associated with transcriptional regulation in lymphomas go by far beyond the usually studied promoter methylation. Furthermore, we characterized the DNA methylome of the mitochondria and the Epstein-Barr virus, whereas upregulation of the latter has already been identified in DAUDI before. As the DAUDI cell line is used over decades in many laboratories throughout the world, the obtained methylome data prove valuable as a "reference epigenome" for future studies.