Fig. 4
From: Cancer network activity associated with therapeutic response and synergism
Assessment of CNA-drug IC50 associations. a Distributions of PCCs for cancer cell lines with mutated or wild-type PIK3CA/R1. The drugs contributing to the negative correlation bias for PIK3CA/R1-mutated cell lines are listed. b Left panel, heatmap showing the results of the PCC analysis between the IC50 profiles of the identified drugs (based on differential CNA-IC50 correlations and mutational status) and the expression profiles of proto-oncogenes and tumor suppressors (by microarray probe; the results for nilotinib are marked). The significant (empirical p < 0.05) correlations are shown. Drugs are color-coded according to the corresponding molecular or biological process target. Right panel, results for the EGFR 201983_s_at probe correlation with all drug IC50 values (distribution) or with nilotinib IC50 (brown lane), originally identified as associated with CNA. c Left panel, heatmap showing the unsupervised clustering of PCCs between the IC50 profiles of associated drugs (based on CNA) with the mutational status of proto-oncogenes or tumor suppressors, or both. Right panels, comparison of the observed number of correlated effects (vertical lines) against equivalent random sets of drugs (distributions). The empirical p values are shown. d Left panel, graph showing the targets of CNA-IC50-based drugs that are differentially expressed between mutated and wild-type cancer cell lines for each of the proto-oncogenes or tumor suppressors analyzed. Red and green indicate over-expression and under-expression in the corresponding mutated setting, respectively. Right panels, comparison of the observed number of differentially expressed targets in PTEN or RB1 mutated cell lines (vertical lines) against equivalent random sets of cancer cell lines. The empirical p values are shown. e Left panel, heatmap showing the results of the PCC analysis between the expression of cancer driver TFs and the IC50 profiles of drugs associated (based on CNA) with the mutational status of proto-oncogenes and/or tumor suppressors. Right panel, PCC distributions for cancer driver TFs (excluding CTNNB1 and PTEN) and the rest of human TFs according to TRANSFAC annotations. The p value of the Wilcoxon rank test is shown