By leveraging the flexibility of the linear mixed model, variancePartition can quantify the variation across individuals within each cell type. Since the variance is analyzed within multiple subsets of the data and each sample is only in a single subset, the total variation explained no longer sums to 1 as it does for standard application of variancePartition. Yet the results allow ranking of dimensions of variation based on genome-wide contribution to variance and enables analysis of gene-level results (Additional file 1). This analysis uses the fact that 34 individuals within monocytes have at least 1 technical replicate, while 41 individuals within T-cells have at least 1 technical replicate.
