|
|
Blood, Vol. 108, Issue 2, 662-668, July 15, 2006
Tumor microenvironment and mitotic checkpoint are key factors in the outcome of classic Hodgkin lymphoma
Blood Sánchez-Aguilera et al.
108: 662
Supplemental materials for: Sánchez-Aguilera et al
Files in this Data Supplement:
- Table S1. Description of the genes in Figure 1 (PDF, 200 KB) -
This table provides a summary of relevant information concerning the biological function of the genes whose expression was found to be associated with the outcome of HL patients. The information was primarily obtained from searches in NCBI databases (Entrez Gene, OMIM and PubMed). A set of supplementary references is provided at the end of the same file. Bibliographic citations in the table (indicated as superscripts) refer to the supplementary references, and not to those in the main text.
- Table S2. Data for Figure 1 (XLS, 154 KB)
-
This presents the base 2-logarithm of the Cy3/Cy5 (RNA of interest / reference RNA) fluorescence-intensity ratios for the 156 clones (145 genes) whose expression was found to be associated with the outcome of HL patients. With the purpose of providing a visually informative image, the expression ratios of each gene were normalized against the average expression of the same gene across the 29 HL samples. In those samples where duplicate hybridizations had been performed (cell lines and centroblasts), the ratio values from the two microarray experiments were averaged. This data-set was used as the input for the clustering procedure shown in Figure 1.
For each of the clones, the following additional information is provided: CNIO identification number; unadjusted p value (obtained from a 100,000-permutation test); False Discovery Rate (FDR) value, calculated by the method of Benjamini and Hochberg; gene name and description. L540, L428, KMH2, HDLM2, L1236 are HL-derived cell lines; CB, centroblasts; LN, reactive lymph nodes; HL, Hodgkin lymphomas. The numbering of the HL samples is the same as in Table 1 (main text), and F and U indicate favorable and unfavorable clinical outcome, respectively. Several clones that were not annotated were assigned to known genes after a BLAST search of the clone sequence; these genes are indicated in parentheses in the “Gene Name” column.
- Figure S1. Gene expression profiles of HL samples, reactive lymph nodes, HL-derived cell lines and normal centroblasts (JPG, 506 KB)
-
Unsupervised hierarchical clustering, using the SOTA algorithm, of the complete data set, excluding those profiles for which less than 80% of valid data were available. For each of the cell lines, RNA was obtained from two independent, similarly passaged cultures, and separate microarray hybridizations were performed. A pool of normal tonsil centroblasts from three donors was hybridized in duplicate. The clustering algorithm yields two main branches. The first comprises the cell lines and the centroblasts, which, in turn, are clustered separately. The second branch includes all the tissue samples. Normal lymph nodes are clustered together, but HL samples are not separated according to their clinical outcome. L540, L428, KMH2, HDLM2, L1236 are HL-derived cell lines; CB, centroblasts; LN, reactive lymph nodes; HL, Hodgkin lymphomas. The numbering of the HL samples is the same as in Table 1 (main text), and F and U indicate favorable and unfavorable clinical outcome, respectively.
- Figure S2. Analysis of the relationship between gene expression and outcome in the validation set of patients (JPG, 1.32 MB)
-
This figure includes the data shown in Figure 2 and additionally contains, for each of the eight markers analyzed, a low magnification field of a stained tissue section and Kaplan-Meier curves comparing the DSS of all four quartiles.
|
|
