Rasooly Danielle, Patel Chirag J
Department of Biomedical Informatics, Harvard Medical School, Boston, Massachusetts.
Curr Protoc Hum Genet. 2019 Apr;101(1):e82. doi: 10.1002/cphg.82. Epub 2019 Jan 15.
Mendelian randomization (MR) is defined as the utilization of genetic variants as instrumental variables to assess the causal relationship between an exposure and an outcome. By leveraging genetic polymorphisms as proxy for an exposure, the causal effect of an exposure on an outcome can be assessed while addressing susceptibility to biases prone to conventional observational studies, including confounding and reverse causation, where the outcome causes the exposure. Analogous to a randomized controlled trial where patients are randomly assigned to subgroups based on different treatments, in an MR analysis, the random allocation of alleles during meiosis from parent to offspring assigns individuals to different subgroups based on genetic variants. Recent methods use summary statistics from genome-wide association studies to perform MR, bypassing the need for individual-level data. Here, we provide a straightforward protocol for using summary-level data to perform MR and provide guidance for utilizing available software. © 2019 by John Wiley & Sons, Inc.
孟德尔随机化(MR)的定义是利用基因变异作为工具变量来评估暴露因素与结局之间的因果关系。通过利用基因多态性作为暴露因素的替代指标,可以在解决传统观察性研究中容易出现的偏倚敏感性问题(包括混杂和反向因果关系,即结局导致暴露)的同时,评估暴露因素对结局的因果效应。类似于随机对照试验中患者根据不同治疗被随机分配到亚组,在MR分析中,减数分裂过程中从亲代到子代的等位基因随机分配会根据基因变异将个体分配到不同亚组。最近的方法使用全基因组关联研究的汇总统计数据来进行MR,从而无需个体水平的数据。在此,我们提供了一个使用汇总水平数据进行MR的简单方案,并为利用现有软件提供指导。© 2019约翰威立国际出版公司