ABSTRACT: We superimposed array comparative genomic hybridization (aCGH) data onto existing expression array profiles (GSE3892) to sift out the pivotal genes underlying the pathogenesis of a well defined series of diffuse large B-cell lymphomas (DLBCL). 280 gained and over-expressed genes were identified, located on chromosomes 1q, 3q, 7, 12p12, 18q and 21q. At the most frequently gained region the oncogene MDMX/MDM4 was concurrently over-expressed, which is a critical regulator of p53 function. In regions of frequent chromosomal loss, 36 genes were concurrently down-regulated, restricted to 6q and 15p. These putative tumor suppressor genes included PERP, MAP3K5, CCNDBP1 and β2M, and loss of 6q was associated with poor clinical outcome. Genes identified and chromosomal regions validated in an independent series of DLBCL patients. Overall design fields: Copy number profiles (array CGH) of a series of 42 Diffuse large B-cell lymphoma patient samples were analyzed a) for frequency of aberrations and b) for impact on gene expression. No duplicate samples or dye-swaps were analyzed. a) Statistical analysis of Array CGH data Calling: Areas of gains and losses and their frequencies throughout the dataset were determined from the Bluefuse ratios within the statistical package â??CGHcallâ?? (van de Wiel et al., 2007). Briefly, after smoothing outliers and median normalization, â??DNAcopyâ?? segmentation (Olshen et al., 2004) was applied within CGHcall, after which a call was assigned to each position on the genome. Within CGHcall the probabilities of the calls were determined per chromosomal arm, and a cellularity correction of 0.75 was applied to each sample because tumor cellularity in all DLBCL is estimated to be maximal 75%, taking into account all stromal, endothelial en infiltrating lymphocytes. After calling, we applied â??CGHregionsâ?? (van de Wiel and van Wieringen 2007) to determine regions of consecutive clones with highly similar calls for all samples. This dimension reduction further increases robustness and statistical power for downstream statistical analyses. Called data were used for all downstream analyses. b) Integration with expression data To integrate called array CGH data with expression array data, the array CGH and expression integration tool, ACE-it (van Wieringen et al., 2006), was used. For ACE-it analysis chromosomal amplifications were included in the gains. ACE-it uses the one-sided Wilcoxon rank sum to test which chromosomal aberrations recurrently affect RNA expression and adjust p-values for multiple testing. Only those clones were taken into account which had at least 20% (9) of the samples in one of the other calling states (gain or loss). Only genes were considered relevant with a false discovery rate of <0.1. References - Olshen AB, Venkatraman ES, Lucito R, Wigler M (2004) Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics 5: 557-72. - van de Wiel MA, Kim KI, Vosse SJ, van Wieringen WN, Wilting SM, Ylstra B (2007) CGHcall: calling aberrations for array CGH tumor profiles. Bioinformatics 23: 892-4. - van de Wiel MA, van Wieringen WN. CGHregions: dimension reduction for array CGH data with minimal information loss. Cancer Informatics. In press 2007. - van Wieringen WN, Belien JA, Vosse SJ, Achame EM, Ylstra B (2006) ACE-it: a tool for genome-wide integration of gene dosage and RNA expression data. Bioinformatics 22: 1919-20.
