The recent increase in obesity levels across many countries is likely to be driven by nongenetic factors. The epigenetic modification DNA methylation (DNAm) may help to explore this, as it is sensitive to both genetic and environmental exposures. While the relationship between DNAm and body-fat traits has been extensively studied, there is limited literature on the shared associations of DNAm variation across such traits. Akin to genetic correlation estimates, here, we introduce an approach to evaluate the similarities in DNAm associations between traits: DNAm correlations. As DNAm can be both a cause and consequence of complex traits, DNAm correlations have the potential to provide insights into trait relationships above that currently obtained from genetic and phenotypic correlations. Utilizing 7,519 unrelated individuals from Generation Scotland with DNAm from the EPIC array, we calculated DNAm correlations between body-fat- and adiposity-related traits by using the bivariate OREML framework in the OSCA software. For each trait, we also estimated the shared contribution of DNAm between sexes. We identified strong, positive DNAm correlations between each of the body-fat traits (BMI, body-fat percentage, and waist-to-hip ratio, ranging from 0.96 to 1.00), finding larger associations than those identified by genetic and phenotypic correlations. We identified a significant deviation from 1 in the DNAm correlations for BMI between males and females, with sex-specific DNAm changes associated with BMI identified at eight DNAm probes. Employing genome-wide DNAm correlations to evaluate the similarities in the associations of DNAm with complex traits has provided insight into obesity-related traits beyond that provided by genetic correlations.

最近许多国家肥胖水平的增加可能是由非遗传因素驱动的。表观遗传修饰 DNA 甲基化 (DNAm) 可能有助于探索这一点，因为它对遗传和环境暴露都很敏感。虽然 DNAm 和体脂性状之间的关系已被广泛研究，但关于这些性状之间 DNAm 变异的共同关联的文献有限。类似于遗传相关性估计，在这里，我们引入了一种评估性状之间 DNAm 关联相似性的方法：DNAm 相关性。由于 DNAm 既可以是复杂性状的原因，也可以是复杂性状的结果，因此 DNAm 相关性有可能提供对性状关系的深入了解，而这些关系目前从遗传和表型相关性中获得。利用来自苏格兰一代的 7,519 名无关个体以及来自 EPIC 阵列的 DNAm，我们通过使用 OSCA 软件中的双变量 OREML 框架计算了身体脂肪和肥胖相关特征之间的 DNAm 相关性。对于每个性状，我们还估计了两性之间 DNAm 的共同贡献。我们确定了每个体脂特征（BMI、体脂百分比和腰臀比，范围从 0.96 到 1.00）之间存在强烈的正 DNAm 相关性，发现比遗传和表型相关性所确定的关联更大。我们发现男性和女性之间 BMI 的 DNAm 相关性与 1 存在显着偏差，在八个 DNAm 探针上发现了与 BMI 相关的性别特异性 DNAm 变化。利用全基因组 DNAm 相关性来评估 DNAm 与复杂性状关联的相似性，除了遗传相关性之外，还提供了对肥胖相关性状的深入了解。