Fig. 1
From: Membrane protein-regulated networks across human cancers
Membrane protein (MP) communities and cancer membrane protein-regulated networks (CaMPNets) across human cancers. a Schematic description of the systematically integrated method (SIM) for predicting protein–protein interactions (PPIs) of 2594 MPs using a combination of template quality (Squl), evolutionary conservation (Ses), sequence similarity (Sjss and Srank), interacting region similarity (Sirs), and network topology (Stopo). b Construction of CaMPNets in human cancers (e.g., breast invasive carcinoma (BRCA), lung squamous cell carcinoma (LUSC), and head and neck squamous cell carcinoma (HNSC)). A CaMPNet consists of an MPP community (an MP with binding partners) and community-regulated pathways. An MPP community-regulated pathway is assessed by the enrichment P values (≤0.05) of co-expressed gene pairs between the differentially expressed genes (up: red and down: green) of an MPP community and a certain pathway based on gene expression profiles in The Cancer Genome Atlas (TCGA) RNA-seq data. c Top, significance (i.e., empirical P value) of the involvement score calculated by 1000 Monte Carlo trials simulated for each community-regulated pathway in individual cancers (for details, see Supplementary Fig. 14a). Bottom, prognostic associations (i.e., adverse or favorable survival) of community-regulated pathways in each cancer type. For Kaplan–Meier curves showing differences in the 10-year overall survival for patients in each community-regulated pathway, an auto-select best cutoff (25–75%) was used, and curve separation was assessed by log-rank test. Clinical outcome data including total 5829 patients in 15 cancer types are assembled from TCGA. HR, hazard ratio
