To investigate multi-ohmic regulation of the metabolome, we used machine learning to predict metabolomic variation across approximately 1,000 different cancer cell lines using matched omics data from eight biomolecular classes. Genomic copy number variation, mutations, DNA methylation, histone post-translational modifications (PTMs), transcriptomics and RNA splice variants, non-coding transcriptomics (miRNA and lncRNA), proteomics, phosphoproteomics. Overall, the metabolome is closely related to the transcriptome, with coding and non-coding RNAs emerging as top predictors. While peripheral metabolites can be predicted by the levels of the corresponding enzymes, central metabolites require combined predictors of signaling and redox pathways and may not reflect the expression of the corresponding pathways. We reconstructed a highly predictable multi-ohmic interaction subnetwork of metabolites, with YAP1 signaling emerging as the top global predictor across four omic layers. We prioritize predictive multi-ohmic features for single-cell and spatial metabolomics assays. Top predictors were enriched for synthetic lethal interactions and synergistic combination therapies targeting compensatory metabolic regulators.
