Figure 7

Network properties. (a) Basic characteristics. For each of the 43 genomes the number of nodes and edges of the four different networks based upon the three methods - chromosome proximity (red), domain fusion (blue), phylogenetic profiling (yellow) - and the combined networks (black) are shown. (b) Giant clusters. The networks are characterized by a giant cluster that relates a dominant fraction of the nodes in the network. In the combined networks of most of the genomes, the giant cluster accounts for 80-90% of the nodes, on average. (c) Universality. The characteristic path distance (global property) and clustering coefficient (local property) of the giant cluster of each network is mapped. Note that networks of the same method tend to cluster together. (d) Degree distributions. The histogram of the number of edges per node is shown on a log-log scale for each network type for Pseudomonas aeruginosa; similar distributions are observed for the other organisms. As in Figure 3, the circles denote values from the observed distributions while the triangles denote the degree distributions of the de novo random networks (see Materials and methods) with the same number of nodes and edges.