Genetic Instruments

We selected genetic variants within 300 kb of the biomarker locus (hereinafter, cis-protein quantitative trait loci [pQTLs]) and used linear regression analyses to estimate the association between each single-nucleotide variant (SNV) and each of the 238 biomarkers. No genetic instrument could be derived for 3 biomarkers, yielding a total of 235 biomarkers investigated through MR. Models were performed separately for each ethnic group, adjusted for age, sex, and the first 5 principal components (body mass index [BMI; calculated as weight in kilograms divided by height in meters squared], type 2 diabetes [T2D], T2D with BMI adjusted, systolic blood pressure [SBP], and type 1 diabetes [T1D]), and then meta-analyzed across ethnicities using fixed-effects models, as previously described.¹⁰ Then, cis-SNVs associated with biomarker concentrations (P < .01) were pruned for linkage disequilibrium at a threshold of r² < 0.1.