Figure 4From: A scalable method for discovering significant subnetworksNormalized frequency of the top 50 most probable size- k subgraphs for each value of k = 6, 7, ..., 20 in Eukaryote-G and Prokaryote-G data sets [15].Back to article page