For each disease, the selected studies were sorted per date of publication regardless of the population of study. Starting for the first study, we built a cumulative database of disease-associated SNPs and their replicability in successive studies. After excluding GWAS with pooled DNAs or focusing on CNVs, each GWAS publication was visually screened for two kinds of association data: the report of a new disease-associated SNPs (discovered SNPs); and the replication status of disease-associated SNPs discovered in previous GWAS (replicated SNPs). In both cases, we recorded three features from each association: (i) Odds Ratio (OR) (ii) confidence interval of the OR and (iii) the p-value.
