Cheng Yu, Ruan Xinjia, Lu Xiaofan, Yang Yuqing, Wang Yuhang, Yan Shangjin, Sun Yuzhe, Yan Fangrong, Jiang Liyun, Liu Tiantian
Research Center of Biostatistics and Computational Pharmacy, China Pharmaceutical University, #639 Longmian Ave, Jiangning District, Nanjing 211100, Jiangsu, China.
Department of Bioinformatics and Computational Biology, The University of Texas, M.D. Anderson Cancer Center, #7007 Bertner Ave, Texas Medical Center, Houston 77030, TX, United States.
Brief Bioinform. 2025 May 1;26(3). doi: 10.1093/bib/bbaf214.
Mendelian randomization (MR) method utilizes genetic variants as instrumental variables to infer the causal effect of an exposure on an outcome. However, the impact of rare variants on traits is often neglected, and traditional MR assumptions can be violated by correlated horizontal pleiotropy (CHP) and uncorrelated horizontal pleiotropy (UHP). To address these issues, we propose a multivariable MR approach, an extension of the standard MR framework: MVMR incorporating Rare variants Accounting for multiple Risk factors and shared horizontal plEiotropy (RARE). In the simulation studies, we demonstrate that RARE effectively detects the causal effects of exposures on outcome with accounting for the impact of rare variants on causal inference. Additionally, we apply RARE to study the effects of high density lipoprotein and low density lipoprotein on type 2 diabetes and coronary atherosclerosis, respectively, thereby illustrating its robustness and effectiveness in real data analysis.
孟德尔随机化（MR）方法利用基因变异作为工具变量来推断暴露因素对结局的因果效应。然而，罕见变异对性状的影响常常被忽视，并且传统的MR假设可能会因相关水平多效性（CHP）和不相关水平多效性（UHP）而被违反。为了解决这些问题，我们提出了一种多变量MR方法，这是标准MR框架的扩展：MVMR纳入了考虑多个风险因素和共享水平多效性的罕见变异（RARE）。在模拟研究中，我们证明了RARE在考虑罕见变异对因果推断的影响的情况下，能够有效地检测暴露因素对结局的因果效应。此外，我们应用RARE分别研究高密度脂蛋白和低密度脂蛋白对2型糖尿病和冠状动脉粥样硬化的影响，从而说明了其在实际数据分析中的稳健性和有效性。
