Westerman Kenneth E, Patel Chirag J, Meigs James B, Chasman Daniel I, Manning Alisa K
Clinical and Translational Epidemiology Unit, Mongan Institute, Massachusetts General Hospital, Boston, MA, United States of America.
Programs in Metabolism and Medical & Population Genetics, Broad Institute of Harvard and MIT, Cambridge, Massachusetts, United States of America.
medRxiv. 2024 Sep 10:2024.09.09.24313018. doi: 10.1101/2024.09.09.24313018.
Discovery and translation of gene-environment interactions (GxEs) influencing clinical outcomes is limited by low statistical power and poor mechanistic understanding. Molecular omics data may help address these limitations, but their incorporation into GxE testing requires principled analytic approaches. We focused on genetic modification of the established mechanistic link between dietary long-chain omega-3 fatty acid (dN3FA) intake, plasma N3FA (pN3FA), and chronic inflammation as measured by high sensitivity CRP (hsCRP). We considered an approach that decomposes the overall genetic effect modification into components upstream and downstream of a molecular mediator to increase the potential to discover gene-N3FA interactions. Simulations demonstrated improved power of the upstream and downstream tests compared to the standard approach when the molecular mediator for many biologically plausible scenarios. The approach was applied in the UK Biobank (N = 188,700) with regression models that used measures of dN3FA (based on fish and fish oil intake), pN3FA (% of total fatty acids measured by nuclear magnetic resonance), and hsCRP. Mediation analysis showed that pN3FA fully mediated the dN3FA-hsCRP main effect relationship. Next, we separately tested modification of the dN3FA-hsCRP ("standard"), dN3FA-pN3FA ("upstream"), and pN3FA-hsCRP ("downstream") associations. The known locus variant rs174535 reached = 1.6×10 in the upstream discovery analysis, with no signal in the downstream analysis ( = 0.94). It would not have been prioritized based on a naïve analysis with dN3FA exposure and hsCRP outcome ( = 0.097), indicating the value of the decomposition approach. Gene-level enrichment testing of the genome-wide results further prioritized two genes from the downstream analysis, and , with links to immune cell counts and function. In summary, a molecular mediator-focused interaction testing approach enhanced statistical power to identify GxEs while homing in on relevant sub-components of the dN3FA-hsCRP pathway.
影响临床结局的基因 - 环境相互作用（GxEs）的发现与转化受到统计功效低和机制理解不足的限制。分子组学数据可能有助于解决这些限制，但将其纳入GxE检测需要有原则的分析方法。我们重点研究了饮食中长链ω - 3脂肪酸（dN3FA）摄入量、血浆N3FA（pN3FA）与通过高敏CRP（hsCRP）测量的慢性炎症之间既定机制联系的基因修饰。我们考虑了一种方法，将整体基因效应修饰分解为分子介导物上游和下游的成分，以增加发现基因 - N3FA相互作用的可能性。模拟结果表明，在许多生物学上合理的情况下，与标准方法相比，上游和下游测试的功效有所提高。该方法应用于英国生物银行（N = 188,700），采用回归模型，该模型使用dN3FA（基于鱼类和鱼油摄入量）、pN3FA（通过核磁共振测量的总脂肪酸百分比）和hsCRP的测量值。中介分析表明，pN3FA完全介导了dN3FA - hsCRP的主效应关系。接下来，我们分别测试了dN3FA - hsCRP（“标准”）、dN3FA - pN3FA（“上游”）和pN3FA - hsCRP（“下游”）关联的修饰。已知的基因座变体rs174535在上游发现分析中达到 = 1.6×10，在下游分析中无信号（ = 0.94）。基于对dN3FA暴露和hsCRP结果的简单分析，它不会被优先考虑（ = 0.097），这表明了分解方法的价值。全基因组结果的基因水平富集测试进一步对下游分析中的两个基因 和 进行了优先排序，它们与免疫细胞计数和功能有关。总之，一种以分子介导物为重点的相互作用测试方法增强了识别GxEs的统计功效，同时锁定了dN3FA - hsCRP途径的相关子成分。
