Chromatin activation profiling of stereotyped CLLs. (A) Study design. ChIP-seq and RNA-seq data are labeled by their source (either previously published or generated in this study). (B) Principal component analysis showing components 1 and 2 in CLL subgroups and normal B cells in 2D and 1D plots. Vertical and horizontal dotted lines respectively point to the cut-off value separating CLLs and normal B cells in component 1 and stereotyped subset 8 from the other samples in component 2. (C) Hierarchical clustering is based on the 1185 robust DRs of subset 8 U-CLL cases as compared with other nonsubset U-CLLs. The M-CLL cases and the normal B-cell subpopulations were added to the clustering to aid in the interpretation of H3K27ac profiles. (D) Boxplots showing the mean H3K27ac of the CLL subgroups and normal B-cell subpopulations in each cluster derived from the hierarchical cluster analysis. (E) Scheme of the strategy used to integrate the differential H3K27ac sites with the expression levels within the same TAD. (F) Barplots based on the integrative analysis showing the percentages of positively correlated regions between H3K27ac and gene expression levels in each cluster derived from the hierarchical cluster analysis.