Fig. 3

Predicting copy number variation using MetOncoFit. The dot plots show the distribution of the values of each feature for the three classes of genes (gain, loss, or neutral). Top 10 features predictive of copy number variation in breast cancer (top panel), NSCLC (middle panel), and melanoma (bottom panel) are shown. The 10-fold cross-validation accuracy is 85%, 94%, and 90% for breast cancer, NSCLC, and melanoma, respectively. Similar set of features were predictive of copy number variation in all three cancers including NCI-60 gene expression levels, catalytic activity, and flux after gene knockout through arginine and proline metabolism, and pyruvate metabolism. In NSCLC, increased flux through the urea cycle/amino group metabolism is associated with a gain in copy number. See Additional file 2: Figure S3 for corresponding data for all nine cancer types. The supplementary website provides data for each gene