Recently, van der Sluis et al. [18] developed a trait-based association test involving the extended Simes procedure (TATES). The TATES calculates a global p-value based on individual p-values of association tests for marginal phenotypes. Specifically, for m-variate phenotypic traits, one can conduct m tests of the association between a candidate SNP and each marginal phenotypic trait and derive mp-values: p1,…pm. Let p(1),…p(m) be the ordered p-values from the smallest to the largest. The Simes multiple procedure declares significance between a SNP and multivariate phenotypic traits at the α level if any of the p-values satisfy p(j)<jα/m [19]. Hence, the global p-value based on the Simes procedure is ptraits= min{mp(j)/j,j=1,…,m}. The TATES improves this procedure by replacing m and j with the effective number of independent traits, me and je, which are estimated from the eigenvalues of the correlation matrix [20, 21]. Since me≤m, this new adjusted global p-value, defined as ptraits= min{mep(j)/je,j=1,…,m} is smaller than the Simes global p-value. Therefore, the TATES is more powerful than the Simes. A simulation study also showed that the TATES is more powerful than
