Table 1.
Motif (core, PWM score cutoffs) | Number of promoter elements found in 9,010 promoter sequences | Number of promoter elements that are conserved in the orthologous mouse promoters | Significance of increase in signal-to-noise ratio (one-tailed)a | ||
Real sequences | Randomized sequences | Real sequences | Randomized sequences | ||
BRE (N/Ab, 0.81) | 2696 (29.9%) | 2503 (27.8%) | 1952 (21.7%) | 1476 (16.4%) | 1.90 × 10-6 |
TATA (0.73, 0.58) | 1848 (20.5%) | 1167 (13.0%) | 1483 (16.5%) | 567 (6.3%) | 1.70 × 10-16 |
INR (0.72, 0.62) | 5949 (66.0%) | 4527 (50.2%) | 5648 (62.7%) | 4040 (44.8%) | 0.02 |
MTE (0.79, 0.53) | 5833 (64.7%) | 5464 (60.6%) | 5123 (56.9%) | 4555 (50.6%) | 0.03 |
DPE (0.92, 0.92) | 1733 (19.2%) | 1650 (18.3%) | 1078 (12.0%) | 695 (7.7%) | 2.90 × 10-11 |
aOne-tailed Fisher's exact test of the real:randomized ratio in columns 4 and 5 versus columns 2 and 3. The p-value in the BRE row, for example, indicates that 1952/1476 is significantly greater than 2696/2503.
bThe "core score" was found to be uninformative for the BRE element, so only the PWM score was used in this case.