EBV-driven B-cell differentiation comprises 4 discrete phases corresponding to the known biology of B-cell differentiation. (A) Hierarchical clustering of normalized publicly available microarray expression data demonstrates the effectiveness of a selected panel of B-cell marker genes in clearly segregating isolated healthy human B-cell subsets into distinct clusters. (B) Schematic outlining EBV WIL in vitro spinoculation infection model. (C) Hierarchical nonsupervised clustering of normalized panel marker gene qRT-PCR expression data from the EBV in vitro infection model results in clear segregation of 4 discrete phases of B-cell differentiation into clusters. Trees demonstrate the similarity of the expression data between samples (below) and genes (left) as determined by the clustering algorithm. Color saturation within a row indicates differences in expression for each gene relative to the average of the sample population; red indicates a higher expression than the mean expression (black), and green indicates a lower expression. Values are quantified by the scale bar that visualizes differences in normalized expression relative to the mean (0). (D) qRT-PCR for Ki-67 was performed at sequential time points to establish the rate of cell proliferation.