For both disease association studies, we performed a standard GWAS using Illumina HumanHap550 BeadChip with over ∼550,000 SNPs, which represent common tagging variants and do not include any of the disease-causing mutations for either condition. We carried out a standard association test on all markers on the chip passing default quality control measures (minor allele frequency >5%, Hardy-Weinberg equilibrium p-value >1×10−6, SNP call rate >95%), using the PLINK software [40]. For the sickle cell anemia GWAS, we compared 194 cases and 7,407 controls of inferred African ancestry via multidimensional scaling, with a genomic control inflation factor of 1.01. For hearing loss, we performed a GWAS on 418 cases and 6,892 control subjects, all of whom were of genetically inferred European ancestry via multidimensional scaling, with a genomic control inflation factor of 1.02.
