Sherlock measures the overlap between data sets that are often genotyped using different platforms. To minimize the chance that informative loci are ignored simply because they are tagged using different SNPs, it is imperative that SNPs in either the GWAS or eQTL data sets (preferably both) be imputed to a larger set. Due to legal/privacy concerns, we do not have access to the genotypes for all eQTL data sets. Therefore, we cannot perform imputation in all cases.