Transferability of type 2 diabetes implicated loci in multi-ethnic cohorts from Southeast Asia

Abstract

Recent large genome-wide association studies (GWAS) have identified multiple loci which harbor genetic variants associated with type 2 diabetes mellitus (T2D), many of which encode proteins not previously suspected to be involved in the pathogenesis of T2D. Most GWAS for T2D have focused on populations of European descent, and GWAS conducted in other populations with different ancestry offer a unique opportunity to study the genetic architecture of T2D. We performed genome-wide association scans for T2D in 3,955 Chinese (2,010 cases, 1,945 controls), 2,034 Malays (794 cases, 1,240 controls), and 2,146 Asian Indians (977 cases, 1,169 controls). In addition to the search for novel variants implicated in T2D, these multi-ethnic cohorts serve to assess the transferability and relevance of the previous findings from European descent populations in the three major ethnic populations of Asia, comprising half of the world's population. Of the SNPs associated with T2D in previous GWAS, only variants at CDKAL1 and HHEX/IDE/KIF11 showed the strongest association with T2D in the meta-analysis including all three ethnic groups. However, consistent direction of effect was observed for many of the other SNPs in our study and in those carried out in European populations. Close examination of the associations at both the CDKAL1 and HHEX/IDE/KIF11 loci provided some evidence of locus and allelic heterogeneity in relation to the associations with T2D. We also detected variation in linkage disequilibrium between populations for most of these loci that have been previously identified. These factors, combined with limited statistical power, may contribute to the failure to detect associations across populations of diverse ethnicity. These findings highlight the value of surveying across diverse racial/ethnic groups towards the fine-mapping efforts for the casual variants and also of the search for variants, which may be population-specific.

Figure 1. Bivariate plots comparing odds ratios observed in each of the ethnic groups with odd ratios established in populations of European ancestry.

(A) Chinese, (B) Malays, (C) Indians, (D) Combined meta-analysis. Each SNP is plotted with a colour that indicates if the SNP was identified through candidate gene studies (black) or linkage studies (red) or candidate-pathway analysis (green) or T2D genome-wide scans (blue).

Figure 2. Regional association plots of the index SNP in CDKAL1.

For each ethnic group, the univariate analysis regional plot, A) Chinese B) Malays C) Indians, is shown together with analysis conditioned on established index SNP rs7754840, D) Chinese E) Malays F) Indians, in populations of Caucasian ancestry. In each panel, the index SNP is represented with a purple diamond and surrounding SNPs coloured based on their r² with the index SNP. Estimated recombination rates reflect the local LD structure in the 500kb buffer around the index SNP and the proxies are plotted on Hapmap values from the Hapmap JPT+CHB. Data for gene annotations are obtained from the RefSeq track of the UCSC Gene Browser (See LocusZoom http://csg.sph.umich.edu/locuszoom/ for more details).

Figure 3. Regional association plots of the index SNP in HHEXX/IDE/KIF11.

For each ethnic group, the univariate analysis regional plot, A) Chinese B) Malays C) Indians, is shown together with analysis conditioned on the index SNP rs6583826, D) Chinese E) Malays, F) Indians, found in meta-analysis across three ethnic groups and established index SNP rs1111875, G) Chinese H) Malays I) Indians, in populations of European ancestry.