Scaling of average clustering coefficients C(k). (I) The primal of E. coli's metabolic hypergraph. (II) Erdös-Rényi random graphs. The cyan dots are C(k) calculated excluding nodes with C = 0. (III) Random graphs with same degree distribution as (I); (IV) primal of rewired versions of E. coli's metabolic hypergraph (see text for more details). For II, III, and IV, the results are based on 100 replica, where the red dots denote the medians of the 100 replica. The red lines are least-squares fitting to C(k) against k using power law. Power coefficients of the fitting are labeled. In each panel, the dashed blue curve corresponds to the points [k, γk], where γk is the smallest C value that a node with degree k can take.
Zhou and Nakhleh BMC Bioinformatics 2011 12:132 doi:10.1186/1471-2105-12-132