High trans-ethnic replicability of GWAS results implies common causal variants

Abstract

Genome-wide association studies (GWAS) have detected many disease associations. However, the reported variants tend to explain small fractions of risk, and there are doubts about issues such as the portability of findings over different ethnic groups or the relative roles of rare versus common variants in the genetic architecture of complex disease. Studying the degree of sharing of disease-associated variants across populations can help in solving these issues. We present a comprehensive survey of GWAS replicability across 28 diseases. Most loci and SNPs discovered in Europeans for these conditions have been extensively replicated using peoples of European and East Asian ancestry, while the replication with individuals of African ancestry is much less common. We found a strong and significant correlation of Odds Ratios across Europeans and East Asians, indicating that underlying causal variants are common and shared between the two ancestries. Moreover, SNPs that failed to replicate in East Asians map into genomic regions where Linkage Disequilibrium patterns differ significantly between populations. Finally, we observed that GWAS with larger sample sizes have detected variants with weaker effects rather than with lower frequencies. Our results indicate that most GWAS results are due to common variants. In addition, the sharing of disease alleles and the high correlation in their effect sizes suggest that most of the underlying causal variants are shared between Europeans and East Asians and that they tend to map close to the associated marker SNPs.

Figure 1. East Asian GWAS find the same risk allele and similar log(OR) than European discovery GWAS.

X axis: log(OR) for the replication stage of the discovery European GWAS. Y axis: log(OR) for the initial stage of East Asian GWAS (Materials and Methods). Dots in blue indicate significant (P<0.05) replication attempts in East Asia; dots in grey indicate non-significant replication attempts. (A) Using all replication attempts; (B) Using only the most-powered replication attempt per SNP; (C) Using the most-powered replication attempt per region. Attempts correspondent to SNPs with MAF<0.01 in East Asians are not shown.

Figure 2. Difference in Average Heterozygosity between Europeans and East Asians.

The x axis represents the distance of each 50-SNP window from the associated SNPs. The y axis shows the average heterozygosity for Europeans minus that of East Asians (SEM indicated in shadow). The band in bisque indicates the windows with significant differences (P<0.01).

Figure 3. Regions harboring non-replicated SNPs present larger differences in LD between Europeans and East Asians.

Measures of difference in LD (varLD scores) are given for sliding windows of 50 SNPs with a 5-SNP step. Measures for replicated and non-replicated SNPs are given as blue and black lines, respectively. Shadowed areas represent the standard error of the mean. The vertical red band indicates that all windows with significant differences (P<0.01) locate in the vicinity of the associated SNPs.

Figure 4. Regions with non-replicated SNPs depart from the genome-wide expectation of regional differences in LD patterns between Europeans and East Asians.

The cumulative distributions of varLD targeted P-values after 1,000 permutations for non-replicated SNPs (n = 64, in gray) and replicated SNPs (n = 47, in blue) are shown. The median (black dashed line) and 95% empirical CI (green area) calculated for 100 groups of 47 SNPs randomly selected from across the genome are also depicted.

Figure 5. Replicabilities against ORs in the discovery study.

For every SNP discovered in Europeans, all the replication attempts in East Asians were considered and classified by bins of OR found in Europeans. The OR of SNPs with risk alleles being major was transformed to ensure OR>1. By means of windows with step 0.3, the average statistical power (empty black circles), average replication success (solid black circles) and effective replicability (the ratio between observed and expected replicability, the two former quantities, red circles) are shown (using only windows with ≥20 attempts). Top values of the graph represent the average date of publication and sample size of discovery GWAS, for bins of 0.1 OR.

Figure 6. Replicabilities against MAFs in the discovery study.

For every SNP discovered in Europeans, all the replication attempts in East Asians were considered and classified by bins of OR found in Europeans. By means of windows with step 0.3, the average statistical power (empty black circles), average replication success (solid black circles) and effective replicability (the ratio between observed and expected replicability, the two former quantities, red circles) are shown (using only windows with ≥20 attempts). Top values of the graph represent the average date of publication and sample size of discovery GWAS, for each bin of MAF.

Figure 7. Similar correlation between European and East Asian log(OR), regardless of the discovery GWAS sample size.

This figure is a redraw of the points shown in Figure 1A but colored according to the discovery GWAS sample size. The same correlations arose when using all replication attempts (as it is shown in Figure) or the filtered (n = 123) set of largest replication attempt per SNP (not shown).