Zhang Jianjun, Sha Qiuying, Hao Han, Zhang Shuanglin, Gao Xiaoyi Raymond, Wang Xuexia
Department of Mathematics, University of North Texas, Denton, Texas, USA.
Department of Mathematical Sciences, Michigan Technological University, Houghton, Michigan, USA.
Hum Hered. 2019;84(4-5):170-196. doi: 10.1159/000506008. Epub 2020 May 16.
The risk of many complex diseases is determined by an interplay of genetic and environmental factors. The examination of gene-environment interactions (G×Es) for multiple traits can yield valuable insights about the etiology of the disease and increase power in detecting disease-associated genes. However, the methods for testing G×Es for multiple traits are very limited.
We developed novel approaches to test G×Es for multiple traits in sequencing association studies. We first perform a transformation of multiple traits by using either principal component analysis or standardization analysis. Then, we detect the effects of G×Es using novel proposed tests: testing the effect of an optimally weighted combination of G×Es (TOW-GE) and/or variable weight TOW-GE (VW-TOW-GE). Finally, we employ Fisher's combination test to combine the p values.
Extensive simulation studies show that the type I error rates of the proposed methods are well controlled. Compared to the interaction sequence kernel association test (ISKAT), TOW-GE is more powerful when there are only rare risk and protective variants; VW-TOW-GE is more powerful when there are both rare and common variants. Both TOW-GE and VW-TOW-GE are robust to directions of effects of causal G×Es. Application to the COPDGene Study demonstrates that our proposed methods are very effective.
Our proposed methods are useful tools in the identification of G×Es for multiple traits. The proposed methods can be used not only to identify G×Es for common variants, but also for rare variants. Therefore, they can be employed in identifying G×Es in both genome-wide association studies and next-generation sequencing data analyses.
许多复杂疾病的风险由遗传和环境因素的相互作用决定。对多个性状的基因-环境相互作用（G×E）进行检测，能够为疾病病因提供有价值的见解，并增强检测疾病相关基因的能力。然而，用于检测多个性状的G×E的方法非常有限。
我们开发了在测序关联研究中检测多个性状的G×E的新方法。我们首先使用主成分分析或标准化分析对多个性状进行转换。然后，我们使用新提出的检验来检测G×E的效应：检验G×E的最优加权组合的效应（TOW-GE）和/或可变权重TOW-GE（VW-TOW-GE）。最后，我们采用Fisher组合检验来合并p值。
广泛的模拟研究表明，所提出方法的I型错误率得到了很好的控制。与交互序列核关联检验（ISKAT）相比，当只有罕见的风险和保护变异时，TOW-GE更具功效；当既有罕见变异又有常见变异时，VW-TOW-GE更具功效。TOW-GE和VW-TOW-GE对因果G×E的效应方向均具有稳健性。在慢性阻塞性肺疾病基因研究中的应用表明，我们提出的方法非常有效。
我们提出的方法是识别多个性状的G×E的有用工具。所提出的方法不仅可用于识别常见变异的G×E，也可用于罕见变异。因此，它们可用于在全基因组关联研究和下一代测序数据分析中识别G×E。
