We used the 1000 Genomes dataset (http://www.1000genomes.org/) as a reference to impute genotyping for our Yale-UPenn and SAGE samples. The program IMPUTE2 was applied for genotyping imputation. EA and AA populations were imputed separately within each dataset.