Figure 3.

Many private and low density peaks are real CEBPA binding events. (A) Heat map of sequence tag density, a one kilobase pair window either side of the center of the peak. The data were then split into two groups, group I (red) contains those peaks that only occurred in one replicate (nā€‰=ā€‰41,229); Group II (black) contains those peaks that occurred in any two replicates (nā€‰=ā€‰81,520). Group I peaks are the 15 subsets of replicate unique peaks ordered by Mmu1_r1 unique peaks (top) to Mmu5_r3 (bottom). Group II peaks are ordered by occupancy. (B)De novo motif analyses using MEME and NestedMica software were able to determine a consensus DNA sequence for CEBPA. (C) Fractions of peaks (as determined by MACS) containing a CEBPA motif were calculated across all replicates and within all peaks using NestedMica for three summit window sizes: 50, 100 and 200 bp. These were compared with multiple iterations of random sequences using the same summit window sizes. (D) For the top 1,000 occupied peaks that occur in multiple replicates (black) and the bottom 1,000 occupied peaks identified in only one replicate (red), both derived consensus motifs were found to be significantly represented in both bins using CentriMo analysis software. A random set of 1,000 genomic intervals of equal size are shown for comparison (orange).

Aldridge et al. Genome Biology 2013 14:R124   doi:10.1186/gb-2013-14-11-r124
Download authors' original image