Copy Number Variation in the Human Genome: Tools, Methods, and Applications to Disease

Human phenotypic variation, including propensities to a wide variety of diseases, is known to have a large heritable component. While the genetic causes of many Mendelian diseases are known, only a small fraction of the proven heritability for common, multifactorial diseases has been explained. This small fraction has mostly been discovered through the use of genome-wide association studies that measure single nucleotide polymorphisms (SNPs). In this thesis, we seek to further the scope of such association studies to include more aspects of human genetic variation.;Copy number variants (CNVs) are classically defined as segments of the genome 1kb to 1Mb in size that are present at variable copy number in different individuals. While first discovered in the genomes of patients with congenital disorders, such variation has recently been found to be present in the genomes of apparently healthy individuals from the general population. It has been proposed that structural variation, including CNVs, may help explain the large component of genetic heritability that has not been accounted for by whole-genome SNP studies. The potential for common CNVs to explain this gap is related to their population genetic properties, while the potential for low-frequency, rare, or private CNVs to explain the gap is dependent their abundance in the human genome.;Here, we successively explore the role of copy number variation across the frequency spectrum and in an integrated fashion with SNPs. First, we develop tools and algorithms required to study SNPs and CNVs side-by-side in a large number of samples. Next, we use these tools to catalogue common copy number polymorphisms (CNPs) in three well-studied populations, and provide the first compendium of CNV genotypes in these populations. We then use this map and set of genotypes to show that CNVs are in linkage disequilibrium with neighboring SNPs, severely limiting the extent to which common copy number variation can explain genetic heritability independent of SNPs. Consistent with these observations, we find that while select CNPs do associate with disease in large cohorts, these associations are generally also captured by associations to nearby SNPs. However, we do discover rare CNVs with large effect sizes in some cohorts including autism, schizophrenia, and tetralogy of Fallot. These CNVs are exceptional in that they are typically recurrent, having independently arisen for most samples, while the vast majority of CNVs found in multiple samples have a shared mutational origin. Together, these studies provide further insight into the genetic architecture of the human species.