Figure 1A shows the cumulative probabilities (P
) for the false positives data set. Poccupied is the probability that a position is occupied by a nucleosome at any point in time. The position of the non-functional binding site is indicated by the red line at sequence position 3000. Cumulative nucleosome occupancy is distributed evenly across the complete length of the sequence, indicating a uniform distribution of nucleosome-positioning sequences. Figure 1B shows the cumulative probabilities for the true positives data set. At the position of the functional binding site (red line) a drop in the cumulative nucleosome occupancy is observed relative to the surrounding sequence, indicating that this area is less likely to be occupied by nucleosomes than the other areas located further up- and downstream. Furthermore, the cumulative probability in this area is comparable to the level observed for the random sequences data set: non-genomic sequences generated by a zero-order Markov model (Figure 1C). Figure 1D shows the normalized cumulative probability in the data set of non-genomic sequences generated by a first order Markov model and is distributed evenly along the whole sequence and at the same level as that in the false positives data set. It is also at the same level as that in the true positives data set at a range of positions distant to the true TFBS (for instance the 1–2 kb region upstream of the TFBS).