PMC:2726282 / 5011-6232 JSONTXT

{"project":"2_test","denotations":[{"id":"18599074-9670816-62517919","span":{"begin":257,"end":259},"obj":"9670816"},{"id":"18599074-10734204-62517920","span":{"begin":843,"end":845},"obj":"10734204"},{"id":"18599074-16772031-62517920","span":{"begin":843,"end":845},"obj":"16772031"},{"id":"18599074-10412974-62517920","span":{"begin":843,"end":845},"obj":"10412974"}],"text":"The majority of the TFs in this network are autoregulated (∼ 70%), of which about two-thirds account for negative loops (see Table 1). This finding is consistent with the results of an analysis performed with a much smaller number of TFs about 10 years ago.17 This predominance of negative autoregulatory loops contrasts with the predominance of positive arcs between different TFs (about 54%, see Table 1). The dominance of positive regulatory interactions in the regulatory network of E. coli is not limited to those among TFs, as when we compute the regulation of all the target genes (3017 arcs) we found that about 54% (1630) are positively regulated, 40% (1206) are repressed, while about 6% (171) are dual regulated. This is especially interesting because a majority of the TFs in bacteria have been reported to act as repressors.12,18,19 The conventions used in Fig. 1 clearly display the hierarchical organisation of the network, with master regulators such as CRP, fumarate and nitrate regulatory protein (FNR) or integration host factor (IHF) each (in)directly regulating a large number of other TFs. Furthermore, the layout emphasises important variations regarding the length of the transcriptional cascades."}