Core GSC SE genes delineate cellular dependencies and define a negative prognostic signature. (a) Bar plot indicating recurrent GSC SE genes detected across GSC models with a frequency >75%. Color code indicates frequency of SE presence. (b) Examples of SE loci regulating CXXC5, SEPT9, SALL3, and SOX2 in GSC, GBM, and NSC samples. (c) Kaplan–Meier survival analysis of the top 250 FASE gene signature scores in TCGA GBM and GBM-low-grade glioma (LGG) RNaseqV2 datasets, respectively. A log-rank test was used to calculate significance of survival differences with P values indicated as *, P ≤ 0.05 and ***, P ≤ 0.0001. Each patient cohort was stratified as high versus low groups at the corresponding median single-sample gene set enrichment analysis z-score. The number of surviving or uncensored patients in each analysis group at noted time points was also displayed. (d) Distribution of top FASE gene signature scores in TCGA GBM low-grade glioma RNaseqV2 dataset among TCGA molecular subtypes and relevant genotypes and epigenotypes (IDH1 mutation; Chr. 1p/19q codeletion; PTEN deletion; PIK3CA mutation; TP53 mutation; MGMT promoter methylation status; TERT promoter methylation status; ATRX mutation status). (e) shRNA knockdown of core SE genes SOX2, SEPT9, CXXC5, CDK6, SALL3, and EGFR in GSC models using two distinct hairpins compared with a nontargeting control. Cell proliferation experiments were performed in biological replicates with more than four technical replicates per time point. Error bars indicated as standard deviation of four technical replicates.