Fig. 4

Biological function characteristics of distinct histone modifier expression patterns. A and B Heatmaps show the results of the gene set variation analysis (GSVA) based on “hallmark gene sets” in the four identified histone modifier expression patterns in the meta-GEO (A) and TCGA-COAD (B) cohorts. Red values represent activated pathways, and blue values represent inhibited pathways. Histone modifier expression patterns are used as sample annotations. C and D Boxplots of GSVA results based on “hallmark gene sets” in the four studied histone modifier expression patterns in the meta-GEO (C) and TCGA-COAD (D) cohorts. Boxes represent 25–75% of values, lines in boxes represent median values, whiskers represent 1.5 interquartile ranges, and black dots represent outliers. Red terms indicate that the corresponding pathway has the highest activation level in patients. E and F Boxplots of OLFM4 high stem cell abundance (left) and mesenchymal cell abundance (right) in the four studied histone modifier expression patterns in the meta-GEO (E) and TCGA-COAD (F) cohorts. Boxes represent 25–75% of values, lines in boxes represent median values, whiskers represent 1.5 interquartile ranges, and black dots represent outliers. G Scatter plots represent the comparison of the protein expression level of the pathway marker genes between patients in the HMC1-3 and HMC4 clusters. H Bar charts summarize the proportions of histone modifier expression patterns in and across different molecular characteristic subgroups. I Sankey diagram of histone modification clusters in groups with different molecular subtypes in the GSE39582 (left) and TCGA-COAD (right) cohorts. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant; Ref, reference; CIMP, CpG island methylator phenotype; CIN, chromosome instability; MT, mutant type; WT, wild type; CMS, consensus molecular subtypes; and TMEC, tumor microenvironment cluster