Figure 2. Weighted gene correlation network analysis (WGCNA) of transcriptome data reveals subtype specific coexpression modules. (A) WGCNA cluster dendrogram groups genes into distinct gene coexpression modules defined by dendrogram branch cutting. (B) Modules-subtype correlation heatmap of three PA subtypes and normal tissues. The cells in the heatmap were colored by the correlation between eigengene expression and each sample group, the correlation coefficients and P values were indicated in each cell. (C–E) KEGG pathway enrichment analysis of genes in the modules significantly associated with different PA subtypes. Adjusted P value < 0.05. (F) Expression profiles of genes in key modules associated with PA subtypes.