Figure 2.

Genome-wide profile of Runx2 occupancy. (A) Distribution of Runx2 binding peaks across the mouse genome were classified into six categories of genomic locations: exon, intron, promoter (-1 kb to +150 bp of TSS), upstream (-1 kb to -20 kb from TSS), TTS region (-150 bp to -1 kb of transcription termination site), and intergenic region. Peak distribution was plotted at each time point by peak number (left panel) and by the percentage of total peaks (right panel). (B) Differential Runx2 enrichment in the six categories of genomic locations compared to the predicted Runx2 motif (see below), to random binding, or to binding of the transcription factor CTCF. Inset provides a magnified view of promoter and TTS regions. (C) The 500 most significant Runx2 peaks, based on MACS [22] significance (P < 1 × 10-10) were used for de novo motif discovery. A Runx2 motif (position weight matrix, top) with strong statistical confidence (P = 4.7 × 10-200) was determined using MEME (MEME suite version 4.7.0) [24]. The known Runx motif (MA0002.2, bottom) in JASPAR [26] was used for comparison using TOMTOM [24]. As shown in (B), the distribution of de novo Runx2 motifs among categories of genomic locations was determined using FIMO [24] at a significance threshold of P < 10-4. (D) Probability plot of the distribution of Runx2 peaks indicating the distances of Runx2 motifs to the peak centers in the top, middle, and bottom third of Runx2 peaks (ranked by MACS scores) versus the probabilities of finding de novo Runx2 motifs at given positions relative to peak center.

Wu et al. Genome Biology 2014 15:R52   doi:10.1186/gb-2014-15-3-r52
Download authors' original image