Fig. 4

Cell-type deconvolution of cancer prognostic gene signatures. A Scaled within-group connectivity of top-ranked cancer prognostic gene sets in different cell-type-specific networks. Cell types presented in more than 10 cancer types were included for analysis. B Weighted degree centrality of the 128 prognostic genes in the fibroblast-specific network in gastric cancer, with emphasis on the top 10 genes. C Weighted degree centrality of genes neighboring the 128 prognostic genes in the fibroblast-specific network in gastric cancer, highlighting the top 10 genes. D Gene set enrichment analysis on the KEGG pathway for the top-ranked 30 genes neighboring the 128 prognostic genes in the fibroblast-specific network in gastric cancer. E Visualization of the network of genes neighboring ITGB1 and several focal adhesion-related genes of the fibroblast-specific network in gastric cancer, with the size of nodes representing the centrality of genes. Brown nodes denote the prognostic genes in the network. F Kaplan–Meier plot for gastric cancer patients of TCGA cohort based on the expression of ITGB1, showcasing the classification of patients into high-expression and low-expression groups by median value for analysis