Genetic Association Test for Multiple Traits at Gene Level

Authors

  • Xiaobo Guo,

    1. Department of Biostatistics, Yale University School of Medicine, New Haven, Connecticut
    2. Department of Statistical Science, School of Mathematics and Computational Science, Sun Yat-Sen University, Guangzhou, China
    Search for more papers by this author
  • Zhifa Liu,

    1. Department of Biostatistics, Yale University School of Medicine, New Haven, Connecticut
    Search for more papers by this author
  • Xueqin Wang,

    1. Department of Statistical Science, School of Mathematics and Computational Science, Sun Yat-Sen University, Guangzhou, China
    2. Zhongshan School of Medicine, Sun Yat-Sen University, Guangzhou, China
    Search for more papers by this author
  • Heping Zhang

    Corresponding author
    • Department of Biostatistics, Yale University School of Medicine, New Haven, Connecticut
    Search for more papers by this author

  • Contact grant sponsor: National Institute on Drug Abuse; Contact grant number: R01 DA016750-09; Contact grant sponsor: NIH Genes, Environment and Health Initiative [GEI]; Contact grant numbers: U01 HG004422 and U01HG004438; Contact grant sponsor: GENEVA Coordinating Center; Contact grant number: U01 HG004446; Contact grant sponsor: Collaborative Study on the Genetics of Alcoholism; Contact grant number: U10 AA008401; Contact grant sponsor: Collaborative Genetic Study of Nicotine Dependence; Contact grant number: P01 CA089392; Contact grant sponsor: Family Study of Cocaine Dependence; Contact grant number: R01 DA013423; Contact grant sponsor: National Institute on Alcohol Abuse and Alcoholism; Contact grant sponsor: National Institute on Drug Abuse; Contact grant sponsor: NIH contract; Contact grant number: HHSN268200782096C.

Correspondence to: Heping Zhang, Department of Biostatistics, Yale University School of Public Health, New Haven, CT 06520. E-mail: heping.zhang@yale.edu

Abstract

Genome-wide association studies (GWASs) at the gene level are commonly used to understand biological mechanisms underlying complex diseases. In general, one response or outcome is used to present a disease of interest in such studies. In this study, we consider a multiple traits association test from the gene level. We propose and examine a class of test statistics that summarizes the association information between single nucleotide polymorphisms (SNPs) and each of the traits. Our simulation studies demonstrate the advantage of gene-based multiple traits association tests when multiple traits share common genes. Using our proposed tests, we reanalyze the dataset from the Study of Addiction: Genetics and Environment (SAGE). Our result validates previous findings while presenting stronger evidence for consideration of multiple traits.

Ancillary