Leclerc Martin, Simard Jacques, Lakhal-Chaieb Lajmi
Département de mathématiques et de statistique, Université Laval, Québec, Canada.
Department of Molecular Medicine, Canada Research Chair in Oncogenetics, Laval University & Genomics Centre, CHU de Québec Research Centre, Québec, Canada.
Genet Epidemiol. 2015 Sep;39(6):406-14. doi: 10.1002/gepi.21914.
In this work, we propose a single nucleotide polymorphism (SNP) set association test for censored phenotypes in the presence of a family-based design. The proposed test is valid for both common and rare variants. A proportional hazards Cox model is specified for the marginal distribution of the trait and the familial dependence is modeled via a Gaussian copula. Censored values are treated as partially missing data and a multiple imputation procedure is proposed in order to compute the test statistics. The P-value is then deduced analytically. The finite-sample empirical properties of the proposed method are evaluated and compared to existing competitors by simulations and its use is illustrated using a breast cancer data set from the Consortium of Investigators of Modifiers of BRCA1 and BRCA2.
在本研究中，我们提出了一种用于在基于家系设计的情况下对删失表型进行单核苷酸多态性（SNP）集关联检验的方法。所提出的检验方法对于常见和罕见变异均有效。针对性状的边际分布指定了比例风险Cox模型，并通过高斯copula对家族依赖性进行建模。删失值被视为部分缺失数据，并提出了一种多重填补程序以计算检验统计量。然后通过解析推导得出P值。通过模拟评估了所提方法的有限样本经验性质，并与现有竞争方法进行了比较，且使用来自BRCA1和BRCA2修饰因子研究联盟的乳腺癌数据集说明了该方法的应用。
