ChIP-seq Analysis

ChIP-seq H3K27ac-defined genomic regions were identified relative to un-immunoprecipitated, fixed chromatin, or “input”, as a negative control. Peaks were called using HOMER with the findPeaks program using the - histone option. Differential peaks between experiments were determined using the getDifferentialPeaks program with default parameters (normalized tag count difference >4 fold and poisson enrichment p-value < 0.0001). Peak merging was performed in HOMER using the “mergePeaks” program to define the union of H3K27ac regions. Each region was centered on the calculated greatest Nucleosome Free Region (NFR) using the -nfr option. The center of the NFR generally has very few H3K27ac tags because this corresponds to the location of TF binding and histone/nucleosome exclusion. In Figure 5B, the central 200 bp sequences of all promoter distal H3K27ac regions (defined as > 3kb from promoter start sites using RefSeq annotations in getDistalPeaks.pl in HOMER) were input for de novo motif analysis using HOMER’s “findMotifsGenome.pl” program. GC-matched 200 bp background sequences are sampled from the genome and used as background for enrichment analysis. Differential H3K27ac regions in cholesterol loading relative to control were defined using “getDifferentialPeaks.pl” with the 24 h cholesterol loaded sample as the foreground and the 24 h control as the background. Motif finding was calculated using the central 100 base pair sequences for cholesterol-up-regulated regions as the foreground and control-up-regulated sequences as the background. Histograms of motif frequency were calculated in annotatePeaks.pl using the -m and -hist options, which output the frequency (y-axes) relative to the center of the NFR (0 bp on x-axes).

(A) The top 5000 most variable H3K27ac gene-distal regions (>3 kb from TSS) across experimental conditions. (B) De novo motif analysis of combined H3K27ac+ regions are shown with predicted factor family TFs, enrichment p-value, and percent occurrence in the data (fgnd) compared to random GC-match genome (bgnd). (C) Frequency of the TEAD motif (y-axis) is plotted relative to the center of H3K27ac+ nucleosome free peak centers (x-axis). (D) The enriched KLF motif is shown as in C. (E) Fosl2/AP-1 and ATF4 motifs are more enriched at H3K27ac+ regions in cholesterol loaded VSMCs relative to the negative (neg) control. (F) The Fosl2 motif is more frequent in cholesterol H3K27ac+ peak centers relative to neg. (G) The ATF4 motif is more frequent in cholesterol H3K27ac+ peak centers relative to neg.