Author:

Keywords:

human genome, copy number alteration, array technologies

Abstract:

In the past decades, advanced array technologies have been developed to look into the molecular structure of the whole human genome in order to unravel human diseases. However, the large amount of data resulting from these innovations have brought new challenges to biologists since the traditional ways to look for disease associated regions or genes are time consuming and inefficient. Due to the high complexity of biological data, proper efficient computational approaches and analysis methods are essential to find correct answers. This thesis is directed towards development of computational approaches for copy number alteration analysis of human genomes. Theoretically, the majority of human genomes contain two copies of chromosomes inherited from each parent. Copy number alteration analysis is the detection of genome regions which contain a copy number different from two. Genome-wide exploration of copy number alteration regions is important since some of these regions are suspected to be relevant to the underlying mechanisms of the diseases. This thesis consists of two parts. The first part describes the development of computational approaches for the detection of copy number alterations in single-cell genomes. Single-cell genome research is a novel area in which the DNA material is extracted from a single cell instead of a pooling of cells as usual. However, the array signals from single-cell genomes often contain a lot of artifacts caused by the amplification of limited amount of DNA material. We develop a novel channel-clone approach to process this challenging single-cell data and correct genome artifacts. By testing this approach in experimental validated genomes, we show that the accuracy of copy number alteration detection is considerably improved after the channel-clone processing. In addition, we apply this approach to single blastomeres extracted from pre-implanted human embryos and discover chromosome instability in thecleavage-stage of human embryos. The second part of this thesis focuses on the analysis of allelic-specific copy number alterations in cancer genomes and the identification of cancer-causing genes. Cancer genomes are the genomes located in cancer cells. Since cancer cells are the result of accumulated mutations, it is straight forward to explore mutated regions within cancer genomes and identify mutated genes residing in these aberrant DNA sections. We extend the previous developed allelic-specific copy number analysis approach (ASCAT) to detect copy number alterations in subclonal populations of cancer cells and apply this method to public available SNP array data. This results in a large consortium of patient-specific copy number profiles from in total 12 cancer types. By comparing copy number profiles, we observe the tumor type specific patterns and identify 89 cancer-related regions including known and candidate cancer-causing genes. Our work reveals a spectacular landscape of cancer genomes across different cancer types and provides valuable directions for future cancer research. In conclusion, this thesis presents different computational approaches to detect aberrations in single-cell genomes as well as cancer genomes and assists biologists and clinicians to create a more comprehensive understanding of human genomes.