KidneyGenAfrica: A multi-cohort Genome-wide association study and polygenic prediction of kidney function in 110,000 continental and diasporan Africans

Abstract

Background: Kidney disease is more prevalent in populations of African ancestry, yet most genome-wide association studies (GWAS) for kidney function markers have been performed in Europeans. To discover additional loci in individuals of African ancestry, we assembled 10 GWASs of the estimated glomerular filtration rate (eGFR) across diverse African regions, including ~26,000 individuals as part of the newly established KidneyGenAfricaconsortium. Additional GWASs of eGFR in ~81,000 African-ancestry individuals in the diaspora were aggregated from the Million Veteran Program (MVP), UK Biobank (UKBB), and the Chronic Kidney Disease Genetics (CKDGen) Consortium. Methods: We performed a three-stage GWAS meta-analysis: (1) Three regional meta-analyses in eastern, western, and southern Africa; (2) a continental African meta-analysis; and (3) a pan-African meta-analysis pooling continental and diaspora studies. We performed fine-mapping, colocalization, functional annotation using MAGMA/FUMA, and a phenome-wide association study (PheWAS). We investigated the role of APOL1 haplotypes in low eGFR in continental Africa. Polygenic scores (PGSs) were estimated from regional, continental, pan-African, and multi-ancestry meta-analyses in a Malawi MEIRU cohort divided into testing and validation sets. Results: The regional meta-analyses identified 28 genome-wide significant loci, including 5 novel loci at FAM72C, LOC645752, OPRM1, KLH1, and LAMA4. The pan-African meta-analysis detected 20 independent loci, including four novel loci (ARG1, OR52H1, TRIM69, and SQRDL). Our fine-mappingidentified four loci with a posterior probability of causality > 0.99. Colocalization recapitulated known eGFR-related genes, and PheWAS showed a pleiotropic profile for 23 of the identified loci, particularly with cardiometabolic, immunological, dermatological, nutritional, and psychiatric traits. The overall APOL1high-risk haplotype frequency in continental Africa was 5%, significantly lower than the approximately 13% observed in African Americans. Notably, we found a limited association between APOL1 variants and low eGFR in continental Africa, which contrasts with the strong APOL1 association with chronic kidney disease observed in African Americans, highlighting distinct genetic risk profiles for kidney disease between African populations and African Americans. PGSs derived from southern African datasets outperformed those from other regional, continental, and multi-ancestry-derived PGSs in the Malawi cohort. Conclusion: We identified novel loci associated with eGFR in individuals of African ancestry from the largest GWAS of eGFR conducted in Africa to date. We observed potential distinct genetic factors that may influence eGFR in continental Africans and African Americans. This suggests that other genetic factors may play a more significant role in eGFR risk among continental Africans. PGSs derived from close genetic distance with discovery cohorts performed better than PGSs derived from other regions, including multi-ancestry data.