In this thesis, I characterize the contribution of rare copy number variation (CNV) to the genetic etiology of Tourette syndrome (TS) and bipolar disorder (BP). I accomplish this using several different study designs and various methods for CNV detection. As array data was widely available for the majority of samples evaluated, I make extensive use of this technology throughout this project and first provide an overview of the technical challenges involved and describe the analytical pipeline I developed to produce reliable CNV calls from such data.
Then, in the largest TS CNV study conducted to date, I report the discovery of the first two genome-wide significant CNVs associated with the disorder, and demonstrate an increased global burden of large, singleton events and CNVs at known, pathogenic loci. Conditioned on this latter observation, I perform an exploratory analysis aimed at gene discovery through the identification of de novo copy number variants from whole-exome sequencing in a sample of affected proband, unaffected parent trios in TS.
I then describe a CNV study of 26 large, multigenerational families with a high incidence of BP from two population isolates, using both microarray and whole-genome sequencing data. While thorough examination of these extended BP pedigrees revealed no segregating variants of large effect, I observe a significant increase in CNV burden across a subset of BP-related genes. Finally, I explore this notion further in a larger North American sample of unrelated individuals ascertained for BP. I demonstrate that although BP cases show no observable differences in the rate or size of rare CNVs, case CNVs affect more highly constrained, brain-expressed genes, and I provide evidence for an increased female CNV burden for BP.