Figure 4.

The probability [p(S > s)] of the scores of the 52 promoters in the training sets from the nine cynaobacteria (cyan), and those of the scores of putative promoters found in the I_U's (pink) and I_C's (blue) and their log odds ratio (red) when both the profiles of NtcA binding sites and −10 like boxes are used for scanning, and the presence of similar sites in the regulatory regions of orthologues in other genomes are also considered. The doted vertical line in each panel shows the largest score cutoff for the genome to include all the binding sites from that genome in the training sets except for PCC6801 and PCC7120, where the largest cutoff is chosen so that only icd of PCC8603, and hetC and icd of PCC7120 are excluded, respectively (see text for details).