Fig. 4
Transcriptomic analysis, selection and clustering of differentially expressed genes. MCF-7 cells were treated with vehicle, 10−9 M E2, 10−5 M glyceollin I and II, or a combination of E2 and each of the glyceollins. Total RNA was extracted, reverse-transcribed, labeled and spotted onto a DNA chip. To select differentially expressed genes, all pairwise conditions were performed. All intensity signals above the overall median and with a fold change ≥2 were selected from each comparison. Then, the selected transcripts were joined and submitted to LIMMA tests. Only transcripts with a p-value ≤0.05 were selected, for a total of 1550 genes (a). These genes were clustered into eight different expression patterns (P1-P8) (b). Each pattern shows the number of genes in parentheses. The conditions are classified according to control, E2, glyceollin I, glyceollin II, E2 and glyceollin I, and E2 and glyceollin II. The most relevant GO terms for each pattern are indicated at the right side of the panel. For example, we found a significant enrichment of genes associated with cell division in P1 (68 genes, 7 expected by chance), as indicated (68/7) in the figure