Hierarchical clustering and PCA of data sets based on cell cycle gene expression. Heat maps in panels A, B, and C show expression levels for the Biocarta cell cycle pathway’s 22 gene components (listed on the right) across samples in the 3 tumor-stage-specific data subsets: SPLs, LPLs, and CRCs, respectively (each containing corresponding samples of normal mucosa, N). Actual sample labels are shown at the top of each heat map (0=normal mucosa; 1=tumor); the groups identified by hierarchical clustering analysis are separated by vertical white lines. (Dendrograms are not shown.) (D) Bi-dimensional projection via PCA of all tumors and normal mucosal specimens using expression levels for the 22 cell cycle-related genes. Each dot represents a tissue sample (pink circle: N; yellow star: SPL; green diamond: LPL; blue square: CRC). The first two components, PC1 and PC2, account for 81% of the variance in this set.