Advanced Visual Analytics Approaches for the Integrative Study of Genomic and Transcriptomic Data

Abstract

The advances in next-generation sequencing (NGS) technology enabled rapid and cost-effective whole genome analyses. Nowadays, it is known that individual organisms have unique genome sequences and that differences between these sequences are the reason for genetic diversity. Furthermore, the biomolecular processes of living organisms are steered by genes and the interplay of their products. Perturbations in these systems often lead to disease. Thus, one of the major question in biomedical research is how genetic variations influence gene function, and how these affect underlying biological pathways and gene interaction networks. One of the most common sources of genetic diversity are single nucleotide variations (SNVs). So-called Genome Wide Association Studies (GWAS) as well as expression Quantitative Trait Locus (eQTL) studies intend to associate SNVs with e.g. disease related binary or quantitative traits. However, available methods are usually limited to statistical analyses and previous approaches to improve the interpretation of the respective results are often insufficient. The goal of this dissertation was the development of new visual analytical approaches to assist purely statistical methods in the identification, characterization and interpretation of SNVs.

Genomic variations, especially SNVs, also play an important role in the immensely growing field of paleogenetics, where DNA of ancient origin is compared to modern DNA with the intention to gain insights into evolutionary history. In this dissertation, a computational pipeline for comparative NGS analyses of ancient and modern DNA samples has been described. Special attention was given to the read merging step, which is required to cope with the quality limitations inherent to ancient DNA (aDNA), in particular DNA fragmentation and nucleotide misincorporation. In addition, aDNA is usually only retrievable in low amounts and it is often contaminated with DNA of modern microorganisms. To solve this issue, a highly economical microarray-based DNA capturing strategy has been developed for the parallel detection and enrichment of aDNA from up to 100 different human pathogens.