Fig. 6

Example of batch effect simulation. A principal component analysis was used to visualize the addition of batch effects to the ‘true expression’ dataset. Each dot represents a subject. In each plot, the y-axis represents PC2 and the x-axis represents PC1 from the principal component analysis of all 10,000 genes included in the microarray gene expression dataset. The panel illustrates the distribution of batches across samples (random) prior to adding batch effects (A), after adding batch effects (B), and again after batch effects were removed using Combat (C)