Comparison of network features.

Modularity , diameter , fraction of nodes in the largest component , global clustering coefficient , and is the average degree of proteins the largest component. ‘Data’ is the empirical data, ‘DUNE’ is the model described here, ‘Vázquez’ is the duplication-only model of [29], ‘Berg’ is the link dynamics model [85], ‘RG’ is random geometric [89], ‘MpK’ is the physical desolvation model presented in [52], and ‘ER’ is an Erdös-Rényi random graph [90]. Simulated values are the median ( standard deviation) over 50 simulations. (See SI for details of each model’s setup and optimization.).