We compared the power of the methods under empirically fixed type I errors to one another and to standard univariate (UV) analysis, univariate analysis of the first principal component of the traits (UV-PCA), and meta-analysis of univariate results (UV-MA). In UV-PCA, the first principal component of a standard principal component analysis is extracted and used in a univariate analysis. In UV-MA, p-values obtained in standard UV GWAS analyses are combined in a meta-analysis approach. Our goal was to provide researchers with insights that will guide the application of the methods to real data.
