Figure 1. Gene expression patterns related to the tissue of origin of the cell lines. Two-dimensional hierarchical clustering was applied to expression data from a set of 1,161 cDNAs measured across 64 conditions. The 1,161 cDNAs were those, of the 9,703 total, that showed the greatest variance in expression level across the 60 cell lines. Data from 64 hybridizations were used, one for each cell line plus the two additional independent representations of each of the cell lines K562 and MCF7. The two cell lines represented in triplicate were correspondingly weighted for the gene clustering so that each of the 60 cell lines contributed equally to the clustering. The cell-line dendrogram is shown in (A), with the terminal branches colored to reflect the ostensible tissue of origin of the cell line (red=leukemia, green=colon, pink=breast, purple=prostate, light blue=lung, orange=ovarian, yellow=renal, grey=CNS, brown=melanoma, black=unknown (NCI/ADR-RES)). A colored representation of the data table, with the rows (genes) and columns (cell lines) in cluster order, is shown in (B). The gene tree is omitted for clarity, but can be obtained at The color in each cell of this table reflects the mean-adjusted expression level of the gene (row) and cell line (column). The color scale used for all the figures in this report is shown in (C). The labels 2A-2D in panel B refer to the clusters of genes shown in detail in Figure 2