Format

Send to:

Choose Destination
See comment in PubMed Commons below
Eur J Hum Genet. 2012 Apr;20(4):469-75. doi: 10.1038/ejhg.2011.219. Epub 2011 Nov 30.

A statistical method for region-based meta-analysis of genome-wide association studies in genetically diverse populations.

Author information

  • 1Department of Epidemiology and Public Health, National University of Singapore, Singapore.

Abstract

Genome-wide association studies (GWAS) have become the preferred experimental design in exploring the genetic etiology of complex human traits and diseases. Standard SNP-based meta-analytic approaches have been utilized to integrate the results from multiple experiments. This fundamentally assumes that the patterns of linkage disequilibrium (LD) between the underlying causal variants and the directly genotyped SNPs are similar across the populations for the same SNPs to emerge with surrogate evidence of disease association. We introduce a novel strategy for assessing regional evidence of phenotypic association that explicitly incorporates the extent of LD in the region. This provides a natural framework for combining evidence from multi-ethnic studies of both dichotomous and quantitative traits that (i) accommodates different patterns of LD, (ii) integrates different genotyping platforms and (iii) allows for the presence of allelic heterogeneity between the populations. Our method can also be generalized to perform gene-based or pathway-based analyses. Applying this method on real GWAS data in type 2 diabetes (T2D) boosted the association evidence in regions well-established for T2D etiology in three diverse South-East Asian populations, as well as identified two novel gene regions and a biologically convincing pathway that are subsequently validated with data from the Wellcome Trust Case Control Consortium.

PMID:
22126751
[PubMed - indexed for MEDLINE]
PMCID:
PMC3306862
Free PMC Article
PubMed Commons home

PubMed Commons

0 comments
How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Nature Publishing Group Icon for PubMed Central
    Loading ...
    Write to the Help Desk