Figure 1: Transcriptional changes and H3K27me3 repositioning during gliomagenesis.

(a) Schematic representation of the experimental procedure followed in this study. Briefly, astrocytes derived from cortices of Ink4a/Arf^−/− pups were infected with the human EGFR* and stereotaxically transplanted into recipient mice. GPCs were derived from PT and re-injected into secondary hosts to assess tumour propagation capability. (b) Schematic representation of the computational pipeline. RNAseq data were used to identify DEGs between Ink4a/Arf^−/− astrocytes overexpressing the human EGFR* (referred to as AstroEGFR*) and GPC (referred to as PT). ChIPseq data defined H3K27me3 target genes in the same cell populations. Integration of the results outlined a subset of genes that are significantly downregulated upon de novo H3K27me3 acquisition in PT. TF motif enrichment analysis (MEA) was performed on DEGs and integrated with RNAseq and ChIPseq data. (c) Left: heatmap of the DEGs between AstroEGFR* and PT. Colours represent Z-scores of the row-wise normalized expression values of log₂ of fragments per kilobase of transcript per million mapped reads (FPKM; red: increased expression, blue: reduced expression). Right: enrichment analysis for canonical pathways in the DEGs. The probability (P-value) of random association between the gene set and given terms is calculated using the right-tailed Fisher’s exact test. Only the top ten most significant pathways are reported; bar represents the −log (P-value). (d) Left: heatmap of normalized tag densities, representing the de novo methylated genes in PT. Each row represents a 6-kb window centred on the gene TSS and extending 3 kb upstream and 3 kb downstream. The signal has been generated by merging the bam files of the biological replicates. Right: enrichment analysis for canonical pathways in the same gene set.