Fig. 5

Construction of KDM-risk score in gastric cancer. A Venn plot reflected 389 KDM phenotype-related genes by overlapping DEGs among three KGRCs. B GO function enrichment of those 389 KDM phenotype-related DEGs. C Overlapping 389 genes with all genes in GSE66229. D Barplot showed the frequency of gene models. E Column plot showed the c-index of KDM_score in TCGA-STAD and GSE66229 cohorts. F Sangi plot illustrated the proportion and distribution of three KGRCs in high and low-risk group. Statistical difference is identified by Kruskal–Wallis H test, * indicating P value < 0.05; ** indicating P value < 0.01; *** indicating P value < 0.001. G-H Expression heatmap of 15 genes in building risk_score and 24 KDM genes was constructed in training set (TCGA-STAD) and validation cohort (GSE66229), respectively