Abstract
BACKGROUND: Serum creatinine (Scr) levels are routinely used to estimate kidney function and health. Individuals of African ancestry have higher Scr levels - controlling for differences in age, sex, size, kidney function, and disease status - compared to individuals from other ancestral backgrounds. The reason for this difference is unknown. We hypothesized that there may be genetic variants found at relatively high frequency in African ancestry groups (African ancestry-enriched variants) that are associated with elevated Scr levels African ancestry individuals. METHODS: Our study sample is made up of participants from the All of Us Research Program. We used whole genome sequence data to estimate genetic ancestry for All of Us participants and selected a cohort of 18,979 participants with two way African-European admixture, available Scr level measures, and demographic covariables. We performed a series of ancestry-informed association studies of Scr levels on this cohort to test our hypothesis of African ancestry-enriched variants associated with Scr. RESULTS: Study participants show an average of 80.8% African and 17.5% European ancestry. Participant Scr levels are positively correlated with African ancestry for females (ρ=0.79) and males (ρ=0.84). The same peak of genome-wide significant associations was identified on chromosome 15 (15q23:45.3Mb-45.5Mb) using standard GWAS, haplotype-based admixture mapping, and ancestry-specific GWAS. The alternate allele for the lead GWAS variant (rs2467850, chr15:45379909:C:T) is positively associated with Scr levels (β=0.07, p=2.28×10(-17)) and found at higher frequency in African (0.413) compared to European ancestry (0.001) groups. Fine mapping identified a credible set of 14 variants co-located with the GATM gene, which encodes a biosynthetic enzyme for creatine, a metabolic precursor of creatinine. 13 of these variants are positively associated with GATM expression, based on a previous study of whole blood eQTL in African Americans, and they all show similar pamerns of African ancestry-enrichment. An Scr polygenic score based on 10 African ancestry-enriched variants completely amenuates the observed association of African ancestry with Scr levels. CONCLUSIONS: Our findings indicate that African ancestry-enriched variants up-regulate the GATM, thereby explaining the higher levels of Scr observed in individuals of African ancestry, and underscore the potential for using genetic data to bemer calibrate kidney function equations.