Table 1  Distributions of positive, negative and dual (auto)regulatory interactions, mean path length16 and observed maximum out- and in-degrees for different sections of the E. coli transcriptional cross-regulatory network shown in Fig. 1
Network section  No. of TFs  Autoregulations  Positive autoregulations  Negative autoregulations  Dual autoregulation  Regulatory arcsa  Positive arcs  Negative arcs  Dual arcs  Average path lengthb  Maximum out-degreec  Maximum in-degreec
All  115  80 (70)  24 (30)  48 (60)  8 (10)  166  90 (54)  67 (40)  9 (6)  2.74  42  9
Carbon sources  24  19 (79)  5 (26)  9 (48)  5 (26)  27  20 (74)  6 (22)  1 (4)  1.53  20  2
Biofilm and motilityd  32  18 (56)  9 (50)  9 (50)  0  52  22 (42)  27 (52)  3 (6)  3.12  8  9
Values in parentheses are percentages. Subnetworks for alternative carbon sources and for biofilm and chemotaxis development processes are defined in Fig. 2a. Note that while the carbon sources module has a relatively high maximum out-degree compared to the biofilm/motility module, the latter has a higher maximum in-degree, clearly suggesting that the motility TF, flhCD (with nine inputs), is directed by several TFs to control its regulation.
a  Regulatory interactions from TFs to others TFs or towards sigma factors.
b  Average path lengths in the (sub)network(s) were calculated with the ViSANT program.16
c  Excluding sigmas.
d  Only the TFs forming cascades ending on biofilm and chemotaxis modules were computed, the autoregulation of CRP was computed in the carbon sources module (Fig. 2a).