Figure 2
From: Development of robust discriminant equations for assessing subtypes of glioblastoma biopsies
Summary of unsupervised and supervised analysis. (A) This figure displays the molecular profile of the 69 probesets with the highest CV and with fluorescence signals higher than 1000, a.u. in at least 30% of cases. Also, these probesets are differentially expressed between the two groups of glioblastomas and their expression values are independent from patient’s gender (see also Supplementary Table 1). Columns are Gb cases and rows probesets. The bottom bar indicates the normalised intensity (arbitrary scale) of probesets per sample. (B) We depict the characteristics of groups (GHE and GLE) detected for each gene signature tested on our local non-standardised data set. Three signatures were evaluated: (1) 69 probesets obtained through our unsupervised approach, (2) Lee’s signature and (3) Colman’s signature, which corresponded to 63 probesets. (C) Features obtained by applying each four probeset-based equation on Lee’s standardised data set. GHE means group of high expression and GLE means group of low expression. Differences between groups in both intensity and age were tested using the Wilcoxon rank-based test, while Pearson's χ2-test was used to evaluate differences in the gender ratio between groups.
