Comparison of network parameters of viral and cellular protein interaction networks

The table shows network parameters for two viral and two cellular protein interaction networks. The SARS network contains 65 interactions between 31 nodes with 6 of those interactions being self-interactions. For comparison purposes, network parameters are also shown for KSHV [26], S. cerevisiae and H. sapiens. Interactions for the cellular networks were derived from the following sources: S. cerevisiae from DIP (the Database of Interacting Proteins) [45], the yeast two-hybrid interactions of H. sapiens from the studies of Stelzl et al. [28] and literature interactions from Rual et al.[27], predicted human interactions (core) from Lehner and Fraser [29] and interactions taken from the HPRD (human protein reference database) [46]. Parameters shown include the number of nodes and edges in the networks, the average degree, characteristic path length (average shortest path), diameter (maximum shortest path), clustering coefficient and for the clustering coefficient enrichment values compared to appropriate random networks (ER and ES). Both types of random networks contain the same number of nodes and edges as the original network. ER networks [47] are created by connecting edges randomly, whereas ES networks are created by an edge swapping strategy which preserves the degree distribution (see [26], [48]). Enrichment values are calculated over the theoretical clustering coefficients of ER networks and the average clustering coefficient of 1000 randomized ES networks.