Figure 2

Frustration index of the 8 biological networks listed in Table 1. In (a) the ratio δ/δ_max is based only on the number of nodes and edges of a network and shows that the frustration index is much lower for transcriptional than for signaling/metabolic networks, see Table 2. The Z-score in (b) takes into account also the topology of a network. Again, the transcriptional networks are more monotone (i.e., less frustrated) than expected from a null model, while metabolic and in particular signaling are less monotone (i.e., more frustrated) than expected. P-values for the Z-score are in Table 2.