Inferring schizophrenia biology from genome-­wide data

Disease loci underlying complex disorders such as schizophrenia(SCZ)are likely to exist as variants of multiple typesbeyond the single nucleotide polymorphisms(SNPs) typically surveyed in genome-­‐wide association studies(GWAS). Large copy number variants (CNV), small insertions and deletions (INDELS),rare compound heterozygous mutations and other classes of genetic variation are all expected to contribute to disease risk. A true understanding of disease genetics cannot be attained without careful consideration of each class of variation and how they are related. Additionally, it is often the case that no or few single variants have strong enough effect sizes to attain the level of statistical significance required to offset the large number of tests performed in a genome-­‐wide single locus survey of any variant type. It has been shown that a cumulative burden of risk alleles is correlated with disease risk, indicating true biological signal within those variants. However, this approach merely implicates that class of variation across the genome and does not refine variants to a subset of risk regions or genes. Defining sets of genes within biological pathways and testing them as a group will allow for both the power of aggregating loci of small effect and interpretation of the underlying biology. This dissertation assesses the role of multiple classes of variation from different technologies in risk to schizophrenia with a particular focus on brain related biological pathways