Skip to main content
Fig. 1 | Journal of Hematology & Oncology

Fig. 1

From: Multiscale protein networks systematically identify aberrant protein interactions and oncogenic regulators in seven cancer types

Fig. 1

Integrative network analysis of pan-cancer protein interactomes. A Data curation. The diagram illustrates omics data types (proteome, transcriptome and mutation) in seven cancer types analyzed in this study, and B Volcano plots of DEPs in tumors. The top 5 up- or down-regulated DEPs in each cancer type are labeled. C. Proteome-specific DEPs: Differential expressions of DEPs in the respective cancer transcriptomes were compared to derive proteome-specific DEPs. The most recurrent proteome-specific DEPs in at least three cancer types were identified by Super Exact Test [11] (Fig. S3D), and they are highlighted in magenta color. D Global protein co-expression networks of seven cancer types. The top network hubs are highlighted and the modules at the resolution of α = 1 are shown as different colored nodes. E Molecular characteristics of the top 10 protein modules in each cancer type. The tracks from the outer most one to the inner most one represent module names (1), cancer type (2), enrichment of the DEP signatures in each cancer type (3, 4), enrichment of the mutational drivers in each cancer type (5), enrichment of the pan-cancer mutational drivers (6), preservation of the protein modules in the respective transcriptomics data (Transcriptome PRV; 7), and preservation of protein modules in the proteomics data of the other cancer types (Cross-cancer PRV; 9–15). There are three scenarios for module preservation: “strong preservation” represented by brown block, “no preservation” by a green block, and “weak preservation” by a grey block. The color intensity bar on the left of the circus plot represents –log10(Fisher’s Exact Test p-value). F Enrichment of the DEP signatures in pan-cancer protein interactomes represented by Pan-cancer protein interaction communites (PCPICs). G Cross-talk among the pan-cancer protein interactomes. In the network, each node represents a PCPIC core and the red and blue links denote positive and negative correlations, respectively. The most enriched pathway for each PCPIC is provided

Back to article page