Personalized Quantitative Metabolic Modeling in the Study of Metabolic Dysfunction after Severe Injury

Metabolic dysfunction that leads to muscle wasting and weakness is a significant clinical problem for patients after severe injury or major surgery. Recent developments of high-throughput "omics" technologies enable the generation of large data sets of clinical samples that can be integrated to build predictive computational models of human metabolism. We applied COBRA to the study of metabolic dysfunctions in the skeleton muscle of patients after severe injury by integrating the knowledge-based human metabolic model of RECON, with the transcriptome, proteome and metabolome data.  To better reflect the quantitative changes of metabolic fluxes seen in patients, we considered the exchange rates of metabolites between human muscle tissue and the blood, and refined the metabolic model by using as constraints the information of enzyme kinetics and the integrated omics data of the individual patient.  This approach of personalized quantitative metabolic modeling (PQMM) was applied to studying 244 skeleton muscle samples from 34 controls and 120 patients, and successfully described known metabolic dysfunctions in the hospitalized patients after severe injury, which include the reduced flux from glycolysis into the TCA cycle, and the increased contributions from the metabolism of amino acids and lipids. In silico gene knock-out and knock-in analyses predicted key genes that increase ATP production ability in patients as potential new targets for intervention.