Ovarian cancer GWAS discovery, expansion and replication (N2 - Project #1)

BRCA 1和BRCA 2的有害突变会大大增加卵巢癌的风险，但这种突变在人群中很少见，它们总共占不到15%的病例。GWAS已经成功地鉴定了多种疾病表型的许多常见易感性等位基因，包括鉴定参与先前未涉及癌症的途径的基因。这对于卵巢癌来说是不可能的，但是最近/即将完成的几个独立的GWAS具有在参与的OCAC研究中复制的潜力，这为实现这一目标提供了机会，并与其他癌症的成功相匹配。在项目1中，第一个目标是从使用类似基因分型平台的四个GWAS中收集数据，使用几种不同的方法来分析数据，并在独立的数据集中进行大规模复制，跟踪最有希望的关联（SNP和CNV）。基因型可从Illumina HumanHap 610芯片获得，通过英国和美国的大约4400例病例和4200例对照。 US GWAS，使用Illumina 317 k芯片对500例病例和300例对照进行分析，以及使用Illumina 1 M珠型芯片对澳大利亚的520例病例和900例对照进行合并DNA分析。此外，英国GWAS已经对另外大约5，000例病例和5，000例对照进行了近22，000个有希望的SNP的基因分型;美国GWAS将在2009年10月对另外6，000例病例和6，000例对照进行16，000个SNP的调查。第二个目标是使用13，680种珠粒类型的定制iSelect Illumina芯片在III期研究人群中复制研究结果，该研究人群包括额外的约3，500例病例和3，500例对照，这些病例和对照未被纳入任何先前的I期或II期基因分型工作。估计三阶段样本量总计约为16，400例病例和17，100例对照，为 检测风险等位基因，对项目2和3至关重要。第三个目的是研究来自所有合并的I期和II期数据的SNP和CNV与卵巢癌生存率的相关性。对于第四个目标，我们将设计并运行具有3，072个SNP的定制iSelect Illumina芯片，用于具有总体存活数据的额外病例的阶段HI研究人群。三阶段生存期样本量估计总数约为13，700例。我们将拟合有序遗传模型来预测已知临床预后因素的总生存率，并对其他主要组织学亚型（例如，粘液和透明细胞）。

Deleterious nnutations in BRCA1 and BRCA2 increase ovarian cancer risk substantially, but such mutations are rare in the population and collectively they account for fewer than 15 percent of cases. GWAS have successfully identified many common susceptibility alleles for multiple disease phenotypes, including the identification of genes involved in pathways not previously implicated in cancer. This has not been possible for ovarian cancer, but the recent/imminent completion of several independent GWAS with the potential for replication in participating OCAC studies provides the opportunity to achieve this and match the success for other cancers, in Project 1, the first aim is tol pool data from four GWAS that used similar genotyping platforms, use several different approaches to analyze the data, and follow-up the most promising associations (SNPs and CNVs) in a large-scale replication in independent data sets. Genotypes are available from the lllumina HumanHap 610 chip on roughly 4400 cases and 4200 controls through UK and US GWAS, on 500 cases and 300 controls using the lllumina 317k chip, and a pooled DNA analysis of 520 cases and 900 controls in Australia which employed the lllumina 1M beadtype chip. In addition, the UK GWAS has genotyped nearly 22,000 promising SNPs on an additional roughly 5,000 cases and 5,000 controls; and the US GWAS will interrogate 16,000 SNPs on an additional 6,000 cases and 6,000 controls in October 2009. The second aim will be to replicate the findings using a custom iSelect lllumina chip of 13,680 beadtypes on a Stage III study population of an additional roughly 3,500 cases and 3,500 controls that have not been included in any of the previous Stage I or Stage II genotyping efforts. The total estimated threestage sample size will be approximately 16,400 cases and 17,100 controls, providing excellent power to detect risk alleles, critical to Projects 2 and 3. The third aim is to examine the association of SNPs and CNVs from all pooled Stage I and Stage II data in relation to ovarian cancer survival. For aim four we will design and run a custom iSelect lllumina chip of 3,072 SNPs on a Stage HI study population of additional cases for whom overall survival data are available. The total estimated three-stage survival sample size will be approximately 13,700 cases. We will fit ordinal genetic models to predict overall survival accounting for known clinical prognostic factors and conduct exploratory analysis of SNP associations within other major histologic subtypes (e.g., mucinous and clear cell).