Figure 6: Zfp423 overexpression differentially affects AstroEGFR* networks.
From: Polycomb dysregulation in gliomagenesis targets a Zfp423-dependent differentiation network

(a) Variable factor map (principal component analysis (PCA)) representing the degree of correlation between AstroEGFR* transcriptomes. The PCA shows that AstroEGFR* cluster according to their responsiveness to Zfp423 overexpression. Correlation coefficients of AstroEGFR*1, AsrtoEGFR*2 and AstroEGFR*3 with the first component are 0.8499869, 0.5120449 and 0.8380024, respectively. (b) Heatmap representing the clusters of genes selectively enriched in the three batches of tumorigenic astrocytes. (c) Pie charts representing the proportion of tumour-retained gene expression for each AstroEGFR*-derived PT. (d) Venn diagrams showing the overlap of Zfp423-dependent genes identified in the three different astrocytes batches upon Zfp423 overexpression. (e) Gene set enrichment analysis showing the enrichment, within the DEGs initially identified between AstroEGFR* and PT, for Zfp423-dependent genes. The interrogation with the Zfp423-dependent genes upregulated in AstroEGFR*2 shows a significant enrichment in the DEGs that are downregulated in the gliomagenic transition from tumorigenic astrocytes to primary gliomas. Black bars represent the position of Zfp423-dpendent genes in the ranked list of DEGs between AstroEGFR* and PT. The green line represents the enrichment score. The nominal P-values associated with the Zfp423-dependent genes upregulated in AstroEGFR*2, the Zfp423-dependent genes upregulated in AstroEGFR*3 and the Zfp423-dependent genes downregulated in AstroEGFR*3 are statically significant (P<0.01).