Next we performed a series of post-GWAS analyses using a protocol outlined in previous analyses [62]. We limited results to the EUR only given the small sample size of the AFR analyses and the lack of AFR predicted transcriptomic expression results in some of the post-GWAS pipelines. To identify functional enrichment, we used MAGMA software (version 1.08), and its recent intersessions (FUMA version 1.3.6) [63], a method for gene-level and gene-set enrichment analysis using GWAS summary statistics. In all the MAGMA-based analyses, SNPs were annotated to the 20,260 coding genes from Ensembl v92, with a 1 kb window for both sides (i.e., start and end). Since GWAS contained EUR and AFR samples, we used the 1000G European and African panels [64] respectively to account for linkage disequilibrium (LD) between SNPs. Finally, we corrected all tests for multiple-testing using a Bonferroni correction.
