variancePartition summarizes the contribution of each variable in terms of the fraction of variation explained (FVE). While the concept of FVE is widely applied to univariate regression by reporting the R 2 value from a simple linear model, variancePartition extends FVE to applications with complex study designs with multiple variables of interest. The linear mixed model framework of variancePartition allows multiple dimensions of variation to be considered jointly in a single model and accommodates discrete variables with a large number of categories. This analysis has a similar motivation as the standard ANOVA method. Yet the linear mixed model framework has several statistical and practical advantages that make it more accurate and generally applicable to complex study designs with multiple dimensions of variation (Additional file 1).
