Fig. 1

Identification of genes highly associated with metabolism and prognosis of ICC using bioinformatics. A: Workflow for the screening process. B: Consensus clustering analysis based on 76 KEGG metabolic pathways from 744 ICC RNA-seq. C: Heatmap of clinical characteristics with survival information. D: Survival difference between the two metabolic clusters (Log-rank test). E: Top 10 differential metabolic pathways between the two clusters. F: Top 20 differential genes between the two metabolic clusters. G: Intersection of differential metabolic genes with TCGA-CHOL and GSE26566. H: Univariate COX regression analysis of 23 metabolic differential genes. I: Expression levels of ANXA1 in TCGA-CHOL (Mann–Whitney U test, ***, p < 0.001) and GSE76297(Wilcoxon matched-pairs test, ***, p < 0.001). J: Kaplan-Meier analysis of overall survival in low and high ANXA1 expression groups in dataset OEP00001105(Log-rank test). K: Top 5 metabolic pathways from the single-gene GSEA enrichment analysis of ANXA1 in the TCGA-CHOL and NODE (OEP001105) datasets