Genome-wide association studies (GWAS) have been very successful in the identification of common genetic variants associated with complex traits. Usually, information on a set of related traits is collected in populations sampled for GWAS. These traits are typically analyzed separately, i.e. in a univariate manner, for association to genome-wide DNA markers. This is often followed by an informal comparison of evidence for association at particular loci across the studied traits (e.g. [1]). However, a joint analysis of multiple, potentially correlated traits, i.e. a multivariate analysis, could be very advantageous for a number of reasons. First, a multivariate analysis has increased power in case of presence of genetic correlation between the different traits; the extra information that is provided by the cross-trait covariance is ignored in univariate analyses [2], [3]. Secondly, most multivariate procedures can perform a single test for association with a set of traits. This reduces the number of performed tests and alleviates the multiple testing burden compared to analyzing all traits separately [2], [4]. Finally, in case of presence of pleiotropy, where a single genetic variant is associated with multiple traits, a multivariate GWAS is more consistent with biology compared to cross-trait comparison of univariate analyses [5].
