Figure 1

Relationship between chromatin marks and eventual transcript levels. (a) Boxplot of the intensities for the ChIP-seq peaks of H3K4me3. Intensities are plotted for the populations with different expression levels measured by RNA-seq and are indicated on the x-axis (Gene Expression Values). Asterisks indicate statistically significant differences, as evaluated by Wilcoxon’s singed rank test (P-values, ***P < 0.001). (b) Scatterplot representing the ChIP-seq peak signal intensities of H3K4me3 on the y-axis and gene expression values on the x-axis (n = 6,105). Pearson’s correlation co-efficiencies of the plots (R = 0.71) are also shown in the graph area. Dotted lines on the x-axis show 10 RPKM and y-axis show 1 × 10⁴ H3K4me3 intensities. Labels on each quadrant of the graph (e.g., ChIP (+) / RNA (−)) are the names given to these set of genes, and are used continually throughout this manuscript. (c, d, e) Graphical representation of the patterns of ChIP-seq (H3K4me3 and pol II) and RNA-seq data for the RSP8 gene (c), PTGS2 gene (d) and NES gene (e). Arrows indicate the direction of transcription and N/A indicates the lack of recognized peaks by MACS. Note that while the ChIP-seq of H3K4me3 and pol II consistently indicate active gene expression of RSP8 and PTGS2 genes, essentially no gene expression was observed for PTGS2 (d). For the NES gene, RNA-seq tags were observed despite the lack of ChIP-seq signals of H3K4me3 and pol II.