Figure 1
From: Insights into cancer severity from biomolecular interaction mechanisms
(a) Methodological overview: histograms show the counts of predicted protein mutation sites at the interface with protein, chemicals and DNA/RNA. (b) Genes mediating the most significantly perturbed interfaces: for each gene, the best q values (FDR) of significantly perturbed, mediated interfaces, is shown. The x-axes and y-axes show q-values obtained from background model 1 and 2. Dots are coloured (grey-black) proportional to the number of unique samples in which a given interface is perturbed, while gene name font size is proportional to the total number of unique samples in which mediated interfaces are significantly perturbed. Dot diameter is proportional to the number of perturbed interfaces. Genes with significantly perturbed interfaces (q <= 0.01 with respect to both background models) in more than 45 unique samples are labelled in green, if they are present in the Cosmic Census or in red if they are not. For space reasons the following significantly perturbed genes (from the Cancer Census) genes are not shown: AKT1, PIK3CA, JAK2, IDH2, IDH1, KRAS, HRAS, PTEN, CHEK2, TP53, GNAS, SMAD4, FBXW7, PPP6C, DICER1, ALK, EGFR, CTNNB1, NRAS, RAC1, SPOP, GNA11.
