Transferability of Tagging-SNPs across disease status: colon cancer and XRCC2

描述(申请人提供)：由于研究设计的不足，特别是遗传变异的选择标准，候选基因的遗传关联研究历来受到影响。现在有几种方法可以表征基因或区域的遗传结构，并识别一组遗传变异或标签SNPs(TSNPs)。这些技术的实施是对复杂疾病进行有效和全面的遗传关联研究的基础。一个仍然存在的重大问题是tSNPs在不同人群中的可转移性。这已被确认为tSNP选择中的一个重要考虑因素，并已涉及到地理血统；例如，作为HapMap项目和NIEHS SNPs计划的组成部分，对四个祖传地理位置进行了调查。可转让性研究才刚刚开始。考虑地理血统的这类研究的最新发现表明，来自HapMap的犹他州tSNP可以在多个欧洲人群中转移。然而，被忽视的一个方面是tSNPs在不同疾病状态下的可转移性。到目前为止，变异发现工作的测序小组是中立的，也就是说，不是基于疾病状态来确定的。因此，假设从中性小组中选择的tSNP将充分代表在随后可能被分析的任何亚群中观察到的单倍型。验证这一假设的研究结果可能会对关联研究设计产生深远的影响。在这个应用中，我们使用XRCC2和结直肠癌状态的真实数据以及模拟数据来研究中性测序面板足以用于tSNP选择这一假设的稳健性。我们将使用来自高危家系的结直肠癌病例和基于人群的研究，根据结直肠癌的肿瘤特征和基因负载，在4个测序小组中表征基因结构。我们的结果将与NIEHS SNPs计划提供的使用中性面板的结果进行比较。选择的XRCC2 tSNP及其单倍型将在三个独立的结直肠资源中被分析与结直肠癌状况和生存率的关联，总共约2,000例病例和约2,000个对照。此外，我们将使用模拟来研究跨疾病祖先的可传递性。我们将通过模拟和分析方法描述中性和基于疾病的小组的相对有效性。特别是，我们将定义使用中性小组的限制，目的是具体解决使用来自HapMap和NIEHS SNPs计划的“现成”数据的潜在限制。

DESCRIPTION (provided by applicant): Genetic association studies of candidate genes have historically suffered due to inadequacies of study design, particularly, the selection criteria for the genetic variants. Several methods now exist to characterize the genetic architecture of a gene or region and identify a set of genetic variants or tagging-SNPs (tSNPs). The implementation of such techniques is the basis for effective and comprehensive genetic association studies for complex diseases. A significant issue that remains is the transferability of tSNPs across populations. This has been acknowledged as an important consideration in tSNP selection and has been addressed for geographic ancestry; for example, four ancestral geographic locations are investigated as constituent parts of the HapMap project and the NIEHS SNPs Program. Transferability studies are just beginning. Recent findings of such studies considering geographic ancestry indicate that the Utah tSNPs from HapMap are transferable across multiple European populations. However, one aspect that has been over-looked is the transferability of tSNPs across disease status. Thus far, sequencing panels for variant discovery efforts are neutral, that is, not ascertained based on disease status. The assumption is, therefore, that tSNPs selected from a neutral panel will adequately represent the haplotypes observed in any subgroup that may subsequently be analyzed. Results from studies testing this assumption have the potential for profound implications on association study design. In this application, we investigate the robustness of the assumption that a neutral sequencing panel is adequate for tSNP selection using real data for XRCC2 and colorectal cancer status as well as simulated data. We will characterize the genetic architecture in 4 sequencing panels based on colorectal cancer tumor characteristics and genetic-loading using colorectal cancer cases from high-risk pedigrees and population- based studies. Our results will be compared to those found using neutral panels, as available from the NIEHS SNPs Program. The XRCC2 tSNPs selected and their haplotypes will be analyzed for association with colorectal cancer status and survival in three independent colorectal resources, totaling approximately 2,000 cases and approximately 2,000 controls. In addition, we will investigate transferability across 'disease' ancestry using simulation. We will describe the relative efficacy of the neutral and disease-based panels across simulation and analysis methods. In particular, we will define the limitations of using a neutral panel with an aim to specifically address the potential limitations of using "off-the-shelf" data from the HapMap and the NIEHS SNPs Program.