Simultaneous testing of multiple genetic variants for association is widely recognized as a valuable complementary approach to single-marker tests. As such, principal component regression (PCR) has been found to have competitive power. We focus on exploring a robust test for an unknown genetic mode of all SNPs, an unknown Hardy-Weinberg equilibrium (HWE) in a population, and a large number of all SNPs. First, we propose a new global test by means of the use of codominant codes for all markers and PCR. The new global test is built on an empirical Bayes-type score statistic for testing marginal associations with each single marker. The new global test gains power by robustly exploiting the Hardy-Weinberg equilibrium in the control population and effectively using linkage disequilibrium among test markers. The new global test reduces to PCR when the genotype for each marker is coded as the number of minor alleles. This connection lends insight into the power of the new global test relative to PCR and some other popular multimarker test methods. Second, we propose a robust test method based on the new global test and the ordinary PCR test built on a prospective score statistic for testing marginal associations with each single marker when the genotype for each marker is coded as the number of minor alleles by taking the minimum value of these two tests. Finally, through extensive simulation studies and analysis of the association between pancreatic cancer and some genes of interest, we show that the proposed robust test method has desirable power and can often identify association signals that may be missed by existing methods.

中文翻译：

---

基于贝叶斯经验得分测验的多基因座基因型数据遗传关联性评估统计方法。

同时测试多种遗传变异的关联性是单标记测试的一种有价值的补充方法。这样，已经发现主成分回归（PCR）具有竞争能力。我们专注于探索针对所有SNP的未知遗传模式，种群中未知的Hardy-Weinberg平衡（HWE）以及大量所有SNP的可靠测试。首先，我们通过对所有标记和PCR使用共密码来提出一项新的全局测试。新的全局测试建立在经验贝叶斯类型得分统计数据的基础上，用于测试与每个单个标记的边际关联。新的全球测试通过在对照人群中大力利用Hardy-Weinberg平衡并有效利用测试标记之间的连锁不平衡来获得动力。当将每个标记的基因型编码为次要等位基因的数目时，新的全局检验简化为PCR。这种联系使我们可以洞悉相对于PCR和其他一些流行的多标记测试方法的新全局测试的功能。其次，我们提出了一种基于新的全局检验和基于前瞻性得分统计量的普通PCR检验的稳健检验方法，用于通过将每个标记的基因型编码为次要等位基因的数量来测试每个标记的边际关联。最低这两个测试的价值。最后，通过广泛的模拟研究和对胰腺癌与某些目标基因之间关联的分析，我们表明，所提出的鲁棒测试方法具有理想的功效，并且通常可以识别现有方法可能会遗漏的关联信号。