Fig. 4.
Hierarchical cluster analysis of myeloma cell lines.
Total RNA from myeloma and nonmyeloma cell lines (5 lymphoma and 1 leukemia) were labeled with Cy5 and hybridized to our 4300 myeloma cDNA microarray together with a Cy3-labeled standard reference pool. All the data were mean centered and clustered using a Pearson correlation coefficient as a similarity metric and average linkage hierarchical clustering.12 Levels of intensity of red squares correlate with the degree of gene expression in cell line experimental samples; conversely, green squares compare the absence of expression among cell lines at a scale relative to the color intensity. Black squares indicate the expression is average among the tested cell lines with ratio corresponding to 1.0 (or log2, ratio of 0), and gray squares represent missing data points. A cluster image representing 2750 of the cDNAs is shown on the left panel. Reproducibility of the array is confirmed by deliberately spotted replicate cDNAs that clustered closely together as shown on the right side of the figure. (A) Immunoglobulin J chain, (B) hypothetical protein MGC3178, (C) immunoglobulin λ light chain, and (D) immunogloblin κ light chain genes are shown on the right panel. (E) Expression of c-maf in different myeloma cell lines based on microarray data. The log2 ratio of fluorescence intensity is shown on the y-axis while myeloma cell lines are shown on the x-axis. The data correlate perfectly with the published literature with the exception of the cell line δ47, which is positive for c-maf expression on the array but is not previously reported to have up-regulated c-mafexpression.