Fig. 1

Gene co-expression network analysis of SnGs in human tissues. a An integrative network biology approach to study SnGs in human tissues. For bulk RNA-seq datasets, a gene co-expression network was constructed in each tissue to identify co-expressed gene modules enriched for the CellAge SnGs and marker genes of different cell types (subpopulations). For each scRNA-seq dataset, unsupervised clustering and differential gene expression analyses were performed to identify cell clusters (subpopulations) and their marker genes. Then the cell-type specific markers were tested for the enrichment in the SnG-enriched modules, revealing cell-type specific SnG signatures in each tissue. The bulk RNA-seq based network analysis and the scRNA-seq based cell type analysis were complementary and cross-validated, providing rich information on co-expression structures and cell-type specificity of SnGs. b Modules enriched for the CellAge SnGs in the 50 tissue-specific gene co-expression networks. Each dot indicates a SnG-enriched module, with the x-axis showing the module size and the y-axis showing the enrichment significance. False Discovery Rate (FDR) in the y-axis was calculated as multiple-testing corrected FET p-value. Color intensity and size of each dot are scaled with the enrichment p-value. The top 20 modules most significantly enriched for the CellAge SnGs are labeled. c The top 20 tissues with the gene modules enriched for the CellAge SnGs. The x-axis shows the number of modules enriched for the CellAge SnGs in each tissue. d The top 20 genes most frequently detected in the SnG-enriched modules. The x-axis indicates the number of enriched modules which contain a given gene. e The top 10 REACTOME pathways for the SnG-enriched modules. The x-axis shows the number of the SnG-enriched modules enriched for a given pathway