Correlations among CTCF occupancy, DNA methylation, and gene expression. (A) PCA analysis of DMCs (left), DEGs (middle), and DBCs (right) based on analyses from EPIC DNA methylation array, RNA-seq, and ChIP-seq for CTCF, respectively. (B) Multiple coinertia analysis of DMCs, DBCs, and DEGs based on EPIC DNA methylation, RNA-seq, and ChIP-seq data, where data from all 3 analyses performed in the same samples were integrated. Each patient is represented by filled circles for DNA methylation data, triangles for ChIP-seq data, and filled squares for gene expression data, while the length of the line that connects the 3 omics types is proportional to the divergence of the data in the same sample. (C) Dot plot of DBCs (AML^all vs NBM) in relation to the change in expression of the related genes comparing AML^all and NBM. The x-axis represents fold change of gene expression, and the y-axis fold change in CTCF occupancy (FDR <0.05). Black dots indicate genes with statistically significant change expression (FDR <0.05, fold change >1.5). χ² test was used to calculate statistical significance between Q1 and Q2 (P < .001). (D) Bar plot showing the genomic distribution of DBCs with differentially expressed gene in panel C. (E) Dot plot of DBCs in relation to fold change of β-values (Δβ-value) for AML^all vs NBM, where the x-axis represents fold change in β-value and the y-axis fold change in CTCF occupancy. Statistical significance is indicated by dashed blue lines for Δβ-value (FDR <0.05, fold change >1.5). Green and red diamonds represent up- or downregulated genes, respectively, while yellow diamonds represent genes with no change in gene expression. χ² test used to calculate statistical significance in gene expression between Q1 and Q4 (P < .001). PC, principal component.