Skip to main content
Fig. 4 | Cancer & Metabolism

Fig. 4

From: Common biochemical properties of metabolic genes recurrently dysregulated in tumors

Fig. 4

MetOncoFit identified topological and dynamic metabolic features that are predictive of gene’s activity on tumor fitness and patient survival. Genes were grouped into three classes based on their expression in poor survival group compared to good survival group—upregulated/neutral/downregulated (see the “Methods” section). The top 10 features predictive of a gene’s impact on patient survival in all three cancers include NCI-60 gene expression levels, catalytic activity, and flux through arginine and proline metabolism after gene knockout. The presence of catalytic activity and expression levels as top predictors suggest that enzyme activity is limiting tumor growth to a greater extent than specific metabolites. Genes that impact flux through arginine, proline, pyrimidine catabolism, and NAD metabolism when knocked out were found to be downregulated in melanoma patients with poor survival (bottom panel). In NSCLC (middle panel), genes that impact flux through pyruvate and the urea cycle/amino group metabolism were upregulated in patients with poor survival. The topological distance from glycogen biomass is negatively correlated with patient mortality in breast cancer (top panel), suggesting that upregulation of enzymes closer to these metabolites is associated with poor survival. The confusion matrices report the accuracy of model predictions using 10-fold cross validation. See Additional file 2: Figure S4 for corresponding data for all nine cancer types. The supplementary website provides data for each gene

Back to article page