Oh Sohee, Huh Iksoo, Lee Seung Yeoun, Park Taesung
* Department of Statistics, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, South Korea.
† Department of Mathematics and Statistics, Sejong University, 209 Neungdong-ro, Gwangjin-gu, Seoul 05006, South Korea.
J Bioinform Comput Biol. 2016 Oct;14(5):1644005. doi: 10.1142/S0219720016440054. Epub 2016 Sep 9.
Most genome-wide association studies (GWAS) have been conducted by focusing on one phenotype of interest for identifying genetic variants associated with common complex phenotypes. However, despite many successful results from GWAS, only a small number of genetic variants tend to be identified and replicated given a very stringent genome-wide significance criterion, and explain only a small fraction of phenotype heritability. In order to improve power by using more information from data, we propose an alternative multivariate approach, which considers multiple related phenotypes simultaneously. We demonstrate through computer simulation that the multivariate approach can improve power for detecting disease-predisposing genetic variants and pleiotropic variants that have simultaneous effects on multiple related phenotypes. We apply the multivariate approach to a GWA dataset of 8,842 Korean individuals genotyped for 327,872 SNPs, and detect novel genetic variants associated with metabolic syndrome related phenotypes. Considering several related phenotype simultaneously, the multivariate approach provides not only more powerful results than the conventional univariate approach but also clue to identify pleiotropic genes that are important to the pathogenesis of many related complex phenotypes.
大多数全基因组关联研究（GWAS）都是通过聚焦于一种感兴趣的表型来识别与常见复杂表型相关的基因变异。然而，尽管GWAS取得了许多成功的结果，但在非常严格的全基因组显著性标准下，往往只能识别和复制少数基因变异，并且这些变异仅解释了一小部分表型遗传性。为了通过利用数据中的更多信息来提高检验效能，我们提出了一种替代的多变量方法，该方法同时考虑多个相关表型。我们通过计算机模拟证明，多变量方法可以提高检测对多种相关表型有同时影响的疾病易感性基因变异和多效性变异的效能。我们将多变量方法应用于一个对8842名韩国个体进行327872个单核苷酸多态性（SNP）基因分型的GWA数据集，并检测与代谢综合征相关表型相关的新基因变异。同时考虑多个相关表型，多变量方法不仅比传统的单变量方法提供了更强有力的结果，还为识别对许多相关复杂表型发病机制重要的多效性基因提供了线索。
