Identifying candidate causal variants via trans-population fine-mapping

Authors

  • Yik-Ying Teo,

    Corresponding author
    1. Department of Statistics and Applied Probability, National University of Singapore, Singapore
    2. Department of Epidemiology and Public Health, National University of Singapore, Singapore
    3. Genome Institute of Singapore, Agency for Science, Technology and Research, Singapore
    • Department of Statistics and Applied Probability, Blk S16, Level 7, 6 Science Drive 2, Faculty of Science, National University of Singapore, Singapore 117546, Singapore
    Search for more papers by this author
  • Rick T.H. Ong,

    1. Genome Institute of Singapore, Agency for Science, Technology and Research, Singapore
    2. Centre for Molecular Epidemiology, National University of Singapore, Singapore
    Search for more papers by this author
  • Xueling Sim,

    1. Centre for Molecular Epidemiology, National University of Singapore, Singapore
    Search for more papers by this author
  • E-Shyong Tai,

    1. Department of Epidemiology and Public Health, National University of Singapore, Singapore
    2. National University Hospital, Singapore
    Search for more papers by this author
  • Kee-Seng Chia

    1. Department of Epidemiology and Public Health, National University of Singapore, Singapore
    2. Centre for Molecular Epidemiology, National University of Singapore, Singapore
    Search for more papers by this author

Abstract

Genome-wide association studies have discovered and confirmed a large number of loci that are implicated with disease susceptibility and severity. Polymorphisms that emerged from these studies are mostly indirectly associated to the phenotype, and the natural progression is to identify the causal variants that are functionally responsible for these association signals. Long stretches of high linkage disequilibrium (LD) benefitted the initial discovery phase in a genome-wide scan, allowing commercial genotyping products with imperfect coverage to detect genomic regions genuinely associated with the phenotype. However, regions of high LD confound the fine-mapping phase, as markers that are perfectly correlated to the causal variants display similar evidence of phenotypic association, hampering the process of differentiating the functional polymorphisms from neighboring surrogates. Here, we explore the potential of integrating information across different populations for narrowing the candidate region that a causal variant resides in, and compare the efficacy of this process of trans-population fine-mapping with the extent of variation in patterns of LD between the populations. In addition, we explore two different strategies for pooling data across multiple populations for the purpose of prioritizing the rankings of the causal variants. Our results clearly establish the benefits of trans-population analysis in reducing the number of possible candidates for the causal variants, particularly in genomic regions displaying strong evidence of inter-population LD variation. Directly integrating the statistical evidence by summing the test statistics outperforms the standard meta-analytic procedure. These findings have direct relevance to the design and analysis of ongoing fine-mapping studies. Genet. Epidemiol. 34: 653-664, 2010.© 2010 Wiley-Liss, Inc.

Ancillary