Extended Data Fig. 7: Immunohistochemistry (IHC) analysis of markers for cell types and pathways in network models.
a-f, Representative IHC images (top), quantification of protein expression by IHC staining (bottom left), and protein expression levels from global proteome data (bottom right) for AGR2 (a), ITGA2 (b), FAP (c), FZD1 (d), PCNA (e), and MCM4 (f) (n = 12, 10, 10, and 13 samples for TS1-4, respectively). H-scores were used for quantification. However, as quantification measures, the percentage of positive cells in each tumor was used for PCNA, and a fraction of positive samples in each TS was used for MCM4. Scale bars = 100 μm. AGR2, P = 3.3 × 10−7 for TS1 vs. other TSs; ITGA2, P = 3.8 × 10−2 for TS2-4 vs. TS1; FAP, P = 2.6 × 10−3 for TS2 vs. other TSs; FZD1 (tumor), P = 8.5 × 10−3 for TS2 vs. other TSs; FZD1 (stroma), P = 1.4 × 10−2 for TS3 vs. other TSs; PCNA, P = 3.2 × 10−6 for TS4 vs. other TSs; and MCM4, P = 2.9 × 10−3 for TS4 vs. other TSs. In violin plots, the line indicates the median value. *, P < 0.05; **, P < 0.01; ***, P < 0.001 by one-sided Student’s t-test (a-e) and Fisher’s exact test (f).
