(Too) great expectations: the challenges in replicating asthma disease genes.
(Too) great expectations: the challenges in replicating asthma disease genes.
复制标题
（太）远大的期望：复制哮喘疾病基因的挑战。
DOI：
10.1164/rccm.200903-0456ed
复制
发表时间：
2009
影响因子：
24.7
通讯作者：
Ober,Carole
中科院分区：
文献类型：
--
作者：
Nicolae,DanL;Ober,Carole
哮喘候选基因的遗传关联研究数量之多可能导致已发表文献中出现许多假阳性报告。评估所报告关联的可靠性的标准方法是在独立样本中复制该关联； “复制”的数量越多，人们就越有信心该基因确实与哮喘易感性有关。另一方面，缺乏重复并不一定意味着原始报告是假阳性，因此解释阴性重复研究更具挑战性。在本期杂志中，Rogers 及其同事（第 1084-1090 页）研究了是否可以使用全基因组关联扫描的数据进一步确认先前涉及的候选标记和/或基因 (1)。这项研究的既定目标是评估先前报道的关联的再现性，并评估这些位点的额外遗传变异的贡献。然而，原则上，这种方法也可用于评估全基因组关联筛查的能力，以检测已知或复制的哮喘易感基因，并评估单个样本中已知基因对哮喘风险的相对贡献。这种方法能在多大程度上解决已报道的 100 多个哮喘相关基因的不确定性 (2) 目前尚不清楚。研究人员评估了儿童哮喘管理计划 (CAMP) 研究中 403 名白人三人组（患有哮喘的儿童及其父母）中 Illumina 550k 阵列中 39 个哮喘候选基因及其周围的基因型数据。这些基因选自截至 2008 年 7 月 1 日的文献综述，并符合以下标准：(1) 至少在两个人群中与哮喘显着相关，(2) 至少一项显着关联研究包括至少 150 个病例和 150 个对照或 150 个三人组，以及 (3) 已发表的关联与单核苷酸多态性 (SNP) 而不是微卫星、SNP 或结构变异。在两阶段设计中，第一阶段称为“SNP 水平复制”，检查阵列上存在的 93 个 SNP，这些 SNP 与先前报道的相关 SNP 相同或具有强连锁不平衡 (LD)（r2> 0.8）；第二阶段称为“基因水平复制”，检查了阵列上的 619 个 SNP，这些 SNP 的次要等位基因频率大于 0.05，这些 SNP 位于 39 个基因的 50 kb 内或之内。第一阶段使用1尾统计检验，第二阶段使用2尾检验。研究人员在第一阶段中发现了与 6 个基因中的 10 个 SNP 之间的适度关联（P 5 0.01-0.05），并在第二阶段中发现了与 15 个基因中的 SNP 相关的证据（P 5 0.002 到 0.05），尽管在调整多重比较后，没有一个“基因水平”关联是显着的。这些结果告诉我们有关哮喘易感基因复制的什么信息？复制遗传关联的主要挑战是获取足够有力的数据集。任何一项研究的功效取决于样本量、相关变体的等位基因频率以及相关变体或基因型的效应大小。前两个组成部分是已知的基因分型标记，但效应大小通常无法准确测量，特别是对于哮喘等受环境影响强烈的疾病。
The sheer number of genetic association studies of asthma candidate genes has likely resulted in many false positive reports in the published literature. The standard approach to assessing the reliability of reported associations is to replicate the association in an independent sample; the greater the number of ‘‘replications’’the more confidence one has that the gene is truly involved in asthma susceptibility. On the other hand, lack of replication does not necessarily imply that the original report was a false positive, and for this reason interpreting negative replication studies is more challenging. In this issue of the Journal, Rogers and colleagues (pp. 1084–1090) investigate whether previously implicated candidate markers and/or genes can be further confirmed using data from a genomewide association scan (1). The stated objectives of this study were to assess the reproducibility of previously reported associations and to evaluate the contribution of additional genetic variation at these loci. However, this approach can also be used, in principle, to evaluate the power of genome-wide association screening to detect known, or replicated, asthma susceptibility genes and to evaluate the relative contribution of known genes to asthma risk in a single sample. The extent to which this approach can address the uncertainty in the greater than 100 asthma-associated genes that have been reported (2) is less clear. The investigators evaluated genotype data from the Illumina 550k array in and around 39 asthma candidate genes in 403 white trios (child with asthma and parents) from the Childhood Asthma Management Program (CAMP) study. The genes were selected from literature reviews through July 1, 2008 and met the following criteria:(1) significant association with asthma in at least two populations,(2) at least one significant association study included at least 150 cases and 150 controls or 150 trios, and (3) the published association is with a single nucleotide polymorphism (SNP) and not a microsatellite, SNP, or structural variant. In a two-phase design, the first phase, called ‘‘SNP level replication,’’examined 93 SNPs that were present on the array and were identical to or had strong linkage disequilibrium (LD)(r2> 0.8) with a previously reported associated SNP; the second phase, called ‘‘gene-level replication,’’examined 619 SNPs on the array with minor allele frequencies greater than 0.05 that were in or within 50 kb of the 39 genes. In the first phase a 1-tail statistical test was used, and in the second phase a 2-tailed test was used. The investigators identified modest associations with 10 SNPs in six genes in the first phase (P 5 0.01–0.05), and evidence for association with SNPs in 15 genes in the second phase (P 5 0.002 to 0.05), although none of the ‘‘gene-level’’associations were significant after adjusting for multiple comparisons. What do these results tell us about replication of asthma susceptibility genes?The main challenge in replicating genetic associations is acquiring a sufficiently powered dataset. The power of any one study depends on sample size, the allele frequency of the associated variant, and the effect size of the associated variant or genotype. The first two components are known for a genotyped marker, but the effect size is generally not accurately measured, especially for diseases with strong environmental influences such as asthma.