Figure 8.

The normalized modularity pattern with respect to the depth of clusters generated. Modularity of a sub-network G′ was calculated by the ratio of the number of edges within G′ to the number of edges starting from the nodes in G′, and it was normalized by the probability of a node in G′ being linked to the other members in G′. As the cluster depth decreases, its normalized modularity has a monotonic increase. It indicates that the modules representing more specific functions have higher modularity.