Multiple genomic regions likely to harbor common genetic variation influencing schizophrenia (SCH) risk have been identified by genome-wide association studies (GWAS) and replicated in multiple independent samples by mega-analysis. Rare sequence variants remain relatively understudied in schizophrenia, but are increasingly detected in excess in cases across a range of common diseases, both in loci known to cause familial forms of disease and more recently in GWAS- identified genes (e.g., for hypertriglyceridemia and Crohn's disease). Although common variants are expected to represent the majority of population risk, understanding rare variants is of great potential value because they may have substantially larger effect sizes and/or more obvious functional significance. The Crohn's study identified variants using an efficient, cost-effective pooled sequencing strategy, which we propose to use here to screen the most implicated genes for variation not tracked by LD with common variants. This is a critical improvement in targeted sequencing designs, because capture, library and sequencing costs, though falling, remain high for the sample N required to achieve desired power. I hypothesize that the specific loci identified by GWAS of schizophrenia also harbor a detectable excess of rare functional sequence variation in cases compared to controls that can be reliably detected by deep, pooled resequencing of these target loci in our case/control sample. To test this hypothesis, I will 1) characterize functional variation (in coding exons, 5'- and 3'- UTRs, promoters, splice sites and any additional regions of evolutionarily constrained sequence not already included) in genes within the GWAS-targeted intervals in 1000 schizophrenia cases and 1000 population controls from Ireland;2) identify specific genetic changes influencing SCH risk through bioinformatic and statistical analyses, and 3) model the impact of sequence variation using forward time population genetic simulations of realistic schizophrenia disease. The study has potential to provide several valuable results to the field, including 1) identification of rare functional variants, 2) insight into the function and dysfunction of specific genes and the underlying pathophysiology of schizophrenia, 3) independent evidence that the targeted genes influences SCH risk, since rare variant signals are unrelated to common SNP LD, 4) improvement in GWAS signals by accounting for unidentified rare variants in the sample and 5) where GWAS signals lie in LD regions containing multiple genes, rare variants may identify the specific liability gene within the regions. Given the dramatic improvement in cost and efficiency provided by the pooled approach, I argue that resequencing these robustly supported target loci is the next logical step to further our understanding of the genetic component of SCH.
Schizophrenia is a common (population frequency ~1%) and often severe mental illness that affects substantial numbers of individuals in the US and worldwide and is a major public health burden because of the early onset, non-fatal course and need for long-term care for many patients. In order to improve understanding of how specific genes are perturbed in schizophrenia, we seek to completely identify rare functional genetic variation (with larger effect sizes and impact on function) in a number of robustly supported, genome-wide significant regions identified by genome-wide association studies (GWAS) through efficient targeted sequencing in pooled samples. The long-term goal is to provide more complete identification of the specific genes impacting schizophrenia risk and better understanding of the kinds of genetic changes that increase risk for schizophrenia, information critical for the development of rational therapy.