Fig. 3: Region-specific biomarker identification and validation.
a Spearman Rank correlation between histomorphological transcriptomic (Ivy GAP) and our own proteomic profiles, leading-edge (LE), infiltrating tumor (IT), cellular tumor (CT), palisading cells around necrosis (PAN), and microvascular proliferations (MVP). Only genes that were identified in both datasets were included in the analysis. b Differential expression matrix analysis based on genes identified based on all samples based on a two-sided t-test at p < 0.01 and 76 degrees of freedom. Values along the diagonal are the number of genes that are differentially abundant against all other regions. c Differential expression matrix comparison by Venn diagram across transcriptionally (Ivy GAP) and proteomically significant biomarkers. Only genes that were identified in both datasets were included. d Regional enrichment of AKAP12 by boxplot highlights increased expression within PAN (n = 154). Data are presented as median values ± IQR and min/max values (whiskers). e H&E image of GBM highlighting necrosis, PAN, and CT. f Region-specific validation within an external cohort IHC demonstrates high staining of AKAP12 within PAN. g Regional enrichment of CD276 using boxplot highlights increased expression within MVP (n = 154). Data are presented as median values ± IQR and min/max values (whiskers). Region-specific validation within an external cohort by IHC demonstrates (h) high staining of CD276 within MVP but (i) no staining of vasculature in LE. j Regional enrichment of PTPRZ1 by boxplot highlights increased expression within IT (n = 154). Data are presented as median values ± IQR and min/max values (whiskers). k H&E image of GBM highlighting LE and IT. l Region-specific validation within an external cohort by IHC demonstrates high staining of PTPRZ1 within tumor cells of the IT. Source data are provided as a Source Data file.
