Large scale genetic association studies have identified many trait-associated variants and understanding the role of these variants in downstream regulation of gene-expressions can uncover important mediating biological mechanisms. In this study, we propose Aggregative tRans assoCiation to detect pHenotype specIfic gEne-sets (ARCHIE), as a method to establish links between sets of known genetic variants associated with a trait and sets of co-regulated gene-expressions through trans associations. ARCHIE employs sparse canonical correlation analysis based on summary statistics from trans-eQTL mapping and genotype and expression correlation matrices constructed from external data sources. A resampling based procedure is then used to test for significant trait-specific trans-association patterns in the background of highly polygenic regulation of gene-expression. Simulation studies show that compared to standard trans-eQTL analysis, ARCHIE is better suited to identify “core”-like genes through which effects of many other genes may be mediated and which can explain disease specific patterns of genetic associations. By applying ARCHIE to available trans-eQTL summary statistics reported by the eQTLGen consortium, we identify 71 gene networks which have significant evidence of trans-association with groups of known genetic variants across 29 complex traits. Around half (50.7%) of the selected genes do not have any strong trans-associations and could not have been detected by standard trans-eQTL mapping. We provide further evidence for causal basis of the target genes through a series of follow-up analyses. These results show ARCHIE is a powerful tool for identifying sets of genes whose trans regulation may be related to specific complex traits. The method has potential for broader applications for identification of networks of various types of molecular traits which mediates complex traits genetic associations.

中文翻译：

---

聚合反式 eQTL 分析检测全血中的性状特异性靶基因组

大规模遗传关联研究已经确定了许多与性状相关的变体，了解这些变体在基因表达下游调控中的作用可以揭示重要的介导生物学机制。在这项研究中，我们提出了聚合tRans关联来检测 pHenotype 特定基因集 (ARCHIE)，作为一种在与性状相关的已知遗传变异集和通过反式关联的共同调节基因表达集之间建立联系的方法。ARCHIE 采用基于反式汇总统计的稀疏典型相关分析-从外部数据源构建的eQTL作图和基因型和表达相关矩阵。然后使用基于重采样的程序在基因表达的高度多基因调控的背景下测试显着的性状特异性反式关联模式。模拟研究表明，与标准的 trans-eQTL 分析相比，ARCHIE 更适合识别“核心”样基因，通过这些基因可以介导许多其他基因的影响，并且可以解释遗传关联的疾病特定模式。通过将 ARCHIE 应用于 eQTLGen 联盟报告的可用trans -eQTL 汇总统计数据，我们确定了 71 个具有明显反式证据的基因网络- 与跨 29 个复杂性状的已知遗传变异组相关联。大约一半 (50.7%) 的选定基因没有任何强反式关联，并且无法通过标准反式 eQTL 作图检测到。我们通过一系列后续分析为靶基因的因果基础提供了进一步的证据。这些结果表明，ARCHIE 是一种强大的工具，可用于识别其反式调节可能与特定复杂性状相关的基因组。该方法具有更广泛的应用潜力，可用于识别介导复杂性状遗传关联的各种类型分子性状的网络。