(264b) Detecting the Warburg Effect and Metabolic Dysregulation By Integrative Control and Enrichment Analysis

Metabolic fluxes impart quantitative information on dynamic cell physiology, which translates to improved therapeutic development and bioproduct synthesis. Metabolic fluxes may be inferred from gene expression, enzyme abundance, and stable isotope tracing. However, these approaches have drawbacks. Metabolic flux analysis (MFA) using stable isotopes and mass spectrometry is resource intensive. Gene set enrichment analysis (GSEA) using gene and protein measurements, although less onerous than MFA, lacks accuracy in identifying enriched metabolic pathways. For instance, GSEA is unable to detect the increased flux through fermentative glycolysis in cancers (i.e., the Warburg effect). Here, we developed a framework, termed Metabolic GSEA or MetGSEA, to reliably predict metabolic fluxes using multi-omic data by integrating metabolic control analysis and enrichment analysis. Our approach accounts for the variability of the effect of a gene on the pathway through flux control coefficients to render it physiologically relevant and accurate. Control coefficients, which are defined as the relative change in the pathway flux in response to the relative change in individual enzyme concentrations, combined with proteomic and/or transcriptomic datasets, reveal significantly altered metabolic fluxes. We used the control coefficients for the individual glycolytic steps from experimental flux measurements and successfully identified the Warburg effect (i.e., increased fluxes and significant enrichment scores for glycolysis) across 666 cancer samples covering nineteen cancer types using RNA-sequencing data from the cancer genome atlas (TCGA). The enrichment scores (in the range of –1 to 1) for glycolysis were 0.749 on average for MetGSEA and 0.365 for GSEA, in the direction of favoring cancerous versus healthy tissues. MetGSEA predicts glycolytic flux upregulation in cancerous tissues in the interquartile range of 3 to 7-fold increases compared to healthy tissues. Thus, our integrative framework combining metabolic control analysis and enrichment analysis accurately depicts the hallmark of cancer. We generalized MetGSEA for the genome-scale metabolic networks and made it publicly accessible as a practical tool for metabolic flux prediction from multi-omic datasets enabling flux-based investigations in physiological processes.