Age-dependent genetic effects on body mass index in the UK Biobank: a SNP-level approach

Human traits frequently change with age, but this variability is often overlooked in traditional genome-wide association studies (GWASs). One such trait, for which genetic and environmental influences change with age, is BMI. Using a SNP-by-age interaction (GxA) GWAS (fastGWA-GE), we estimated age-varying genetic effects on BMI in the UK Biobank ( N unrelated = 380,138; N unrelated + related = 451,696). In the full sample, we identified a significant SNP-by-age interaction in the FTO gene, known for its role in weight regulation, as well as in APOE , APOC1 and TOMM40 , which are implicated in lipid metabolism and other processes, highlighting candidates for further investigation into age-related BMI variability. To assess predictive performance, we calculated polygenic risk scores (PRSs) that account for SNP-by-age interactions from the unrelated sample, explaining <0.1% additional variance beyond standard PRSs. Genetic correlations between the unrelated age-dependent summary statistics from various ages (40, 50, 60, and 70) and BMI ranged from 0.9139 to 0.9954, two being significantly different from 1. Taken together, these results suggest that incorporating SNP-by-age interactions in genetic studies offers unique insights into the genetic architecture of traits and their age-varying genetic associations, which can be essential for clarifying the mechanisms underlying age-related diseases.