Figure 5

Posterior probability plots. These plots show the distributions of the posterior probabilities P(S _t = 1), i.e., the probability that there is a binding site at window t. On the left are the plots of the probabilities for the HNF4α and H3K4me3 data sets shown in full scale, and on the right are the same plots with a closer view of the tails of the distributions. We can see that the vast majority of the regions have zero probability of being enriched and the remaining regions have probabilities very close to 1. Only a handful of regions lie in the middle of the range, and these are likely to lie in the tails of large peaks. To include some of these regions, we set the threshold to be equal to a posterior probability of 0.50 and called every region with a probability greater than that as enriched.