GWAS focus on describing new variants across the genome rather than validating the findings from previous GWAS. Instead, many replication attempts consist on genotyping limited sets of SNPs previously discovered by GWAS in independent samples. They tend to be published independently from GWAS and, hence, our replicability database may have failed to gather many replication attempts that occur outside the setting of genome-wide studies. Since endeavoring to analyze all the literature available for the 28 diseases in our database would have required a massive effort, we randomly selected six diseases to address this possibility (Table S10). For each disease, we searched all the publications citing each of the disease-associated variants present in our database, as well as for the original GWAS publications initially describing them.
