Fig. 3From: Identifying network biomarkers of cancer by sample-specific differential networkValidating sample-individual differential networks and predicting driver genes in cancer. A Four distribution samples ranging from 100 to 1000, 1000 to 10,000 and 10,000 to 100,000. B Cross validation of three cancer. C Cross validation of three gastric cancer databases. D The proportion of significant samples in the enrichment analysis of top 100, 50, 30, 20 and 10 highest degree genes for BRCA DSSN in the KEGG pathway and compare with previous method (SSN). E The proportion of significant samples in the enrichment analysis of top 100, 50, 30, 20 and 10 highest degree genes for BRCA DSSN in the CGC database and compare with previous method (SSN). F The proportion of somatic mutation genes to be driver mutation genes in top 100, 50, 30, 20 and 10 highest degree genes for each DSSNBack to article page