A clear prediction can be made if, as our results suggest, a substantial fraction of associations reported by European GWAS are caused by common variants with similar effect sizes across the two ancestral groups: whenever associations were successfully replicated, the frequencies of tagSNPs and causal variants and their LD relationships should be similar in the two groups. In other words, levels of heterozygosity and LD patterns should be more similar between populations in the genomic regions that contain successfully replicated SNPs than in the genomic regions with European-associated SNPs that have not reached significance in East Asians. To test this prediction, we compared the inter-continental similitude of heterozygosity and LD in genomic regions harboring two different groups of disease-associated SNPs: the 47 SNPs discovered in Europeans that have been successfully replicated in every attempt with East Asians and the 65 SNPs that have never been positively replicated.
