Analysis of gene-expression networks using WGCNA. (A) Correlations between consensus modules identified by WGCNA and specific experimental conditions. The y-axis shows the 14 consensus modules that were randomly assigned different colors for identification, and the x-axis shows the 11 experimental conditions defined within the 2 GEP studies.^11,18  The heat map shows the correlation between the expression of each module’s eigengene and experimental condition; cells show correlation coefficient (top) and corresponding P value (bottom). Cells are color-coded using correlation values according to scale on the right. For the Vallat et al study: S, anti-IgM stimulated; US, no stimulation. Time of stimulation (60, 90, 210, and 390 minutes) is indicated. For the Herishanu et al study, the source of the sample is shown. BM, bone marrow; LN, lymph node; PB, peripheral blood. (B) Correlations between kME and fold difference in expression between LN compared with other sites (top) and after anti-IgM stimulation in vitro (bottom) for the 344 genes comprising the blue module. (C) “TreeMap” view of over-represented GO terms in the blue module, generated using REVIGO. Each rectangle represents a cluster of related GO terms. The representatives are joined into “superclusters” of loosely related terms, visualized with different colors. Size of the rectangles reflects Benjamini-Hochberg adjusted P values.