Figure 2

Number of predicted TF binding sites within 200-bp of the TSS compared to background sequences. Each log-scale plot shows the total number of predicted binding sites in the set of 314 200-bp promoter regions on the y-axis. The x-axis represents the number of sites found in the preceding 200-bp regions (upper panels, A and C) or the average number of sites predicted in 10,000 shuffled sequence datasets (lower panels, B and D). Each data point represents a family of transcription factor matrices in the MatBase database; MatInspector analysis was performed using the default parameters (see Additional File 4 for analysis using less stringent parameters, and Table 1 and Additional Files 5 and 6 for full results and matrix family names). Panels A and B (left) show results before masking O/E sites and TATA boxes, and panels C and D (right) show analysis after masking those sites. Data points for which one of the two values is 0 are not plotted (none shows statistically significant enrichment). Solid black symbols represent matrix families showing statistically significant enrichment; solid gray symbols represent matrix families showing statistically significant depletion; the blue square symbol highlights the V$NOLF family of matrices representing the O/E binding site (as expected, no O/E sites were predicted after applying the mask, so O/E does not appear in panels C and D); the red diamond symbol highlights the O$VTBP family of matrices representing TATA boxes.