Finding single-gene copy number variants in schizophrenia

描述(申请人提供)：精神分裂症(SCZ)是一种高度遗传性的精神疾病，其病理生理学难以捉摸，几乎没有新的治疗方法正在开发中。通过分离大量影响风险的精确因果突变，我们对SCZ生物学基础的理解将大大加快。拷贝数变异(CNV)是一个诱人的致病突变来源，因为它们通过改变基因剂量或结构，提供了明确的作用方向和分子机制。我们建议利用由NIMH资助的瑞典精神分裂症协会大规模研究中使用的多种技术和大样本的力量来识别与SCZ相关的基因和外显子水平的CNV。我们的首要目标是发现由特定基因拷贝数变化引起的风险位点，这可能会导致对SCZ的更好理解。首先，我们将结合来自互补技术的多个CNV调用集来识别名义上与SCZ相关的基因和外显子水平的CNV。可获得的数据包括5,001例SCZ病例和6,234名对照的全基因组SNP、外显子组基因分型阵列和外显子组测序。其次，我们将整合所有三个数据集的归一化原始数据来构建一个虚拟数组，并使用专门的隐马尔可夫模型(VAMM)来调查基因CNV，以产生高密度的CNV呼叫。第三，我们将创建一个与SCZ相关的基因/外显子水平CNV的优先列表，并使用精确和高灵敏度的独立方法学验证50个假定的CNV。最后，我们将通过查询现有的来自20,000个病例和20,000个对照的基因组数据来评估有效的CNV区域是否与SCZ无关。到目前为止，通过CNV分析只发现了几个特定的基因，所有这些都为SCZ的病理生理学提供了新的见解。我们认为这些基因只是冰山一角，目前的不足仅仅归因于缺乏以基因为重点的CNV评估方法。如果这项R21研究能够发现一个新的外显子或基因水平的CNV，其剂量改变了SCZ的风险，这将代表着我们在理解SCZ的生物学基础方面取得的重要进展。任何确定的CNV都将为R01应用程序提供基础，以了解其剂量与SCZ风险之间的联系机制。

DESCRIPTION (provided by applicant): Schizophrenia (SCZ) is a highly heritable psychiatric disorder with an elusive pathophysiology and few novel treatments under development. Our understanding of the biological basis of SCZ would be greatly accelerated by the isolation of a substantial number of precise causal mutations that influence risk. Copy number variants (CNVs) are an attractive source for causative mutations since, by altering gene dosage or structure, they provide a clear direction of effect and molecular mechanism. We propose to leverage the power of multiple technologies and large samples used in the Large-Scale Swedish Schizophrenia Association Study funded by NIMH to identify gene- and exon-level CNVs associated with SCZ. Our overarching goal is to discover risk loci resulting from changes in copy number in specific genes that could lead to improved understanding of SCZ. First, we will combine multiple CNV call sets from complimentary technologies to identify gene- and exon-level CNVs nominally associated with SCZ. Data available include genome-wide SNP, exome genotyping arrays and exome sequencing in 5,001 SCZ cases and 6,234 controls. Second, we will integrate the normalized raw data from all three datasets to construct a "virtual array" and use a specialized hidden Markov model (VAMM) to survey genic CNVs to yield high-density CNV calls. Third, we will create a prioritized list of gene/exon level CNV associations with SCZ and validate 50 putative CNVs using a precise and high sensitivity independent methodology. Finally, we will evaluate whether validated CNV regions associate with SCZ beyond chance by querying existing genomic data available from a distinct set of 20,000 cases and 20,000 controls. To date, only a few specific genes have been implicated via CNV analysis, all of which provided novel insights into SCZ pathophysiology. We contend that these genes are the tip of an iceberg and the current deficiency is simply attributed to the lack of methods for gene-focused CNV evaluation. If this R21 study can identify even one new exon- or gene-level CNV whose dosage alters SCZ risk, it will represent an important advance in our understanding of the biological basis of SCZ. Any identified CNV would provide the basis for an R01 application to understand the mechanisms linking its dosage to SCZ risk.