Interpreting Metabolomic Profiles using Unbiased Pathway Models

• DOI: 10.1371/journal.pcbi.1000692
• Publication Date: 2010-03-03T00:11:12Z
• Authors: Rahul C. Deo; Luke Hunter; Gregory D. Lewis; Guillaume Pare; Ramachandran S. V...; Daniel Chasman; Thomas J. Wang; Robert E. Gerszten; Frederick P. Roth
• Keywords: Metabolome; Metabolic network; Metabolic pathway

Human disease is heterogeneous, with similar phenotypes resulting from distinct combinations of genetic and environmental factors. Small-molecule profiling can address heterogeneity by evaluating the underlying biologic state individuals through non-invasive interrogation plasma metabolite levels. We analyzed profiles an oral glucose tolerance test (OGTT) in 50 individuals, 25 normal (NGT) impaired (IGT). Our focus was to elucidate processes. Although we initially found little overlap between changed metabolites preconceived definitions metabolic pathways, use unbiased network approaches identified significant concerted changes. Specifically, derived a edges drawn reactant product nodes individual reactions all substrates enzymes transporters. searched for "active modules"—regions enriched changes Active modules relationships among highlighted importance specific solute carriers profiles. Furthermore, hierarchical clustering principal component analysis demonstrated that OGTT naturally grouped according activities System A L amino acid transporters, osmolyte carrier SLC6A12, mitochondrial aspartate-glutamate transporter SLC25A13. Comparison NGT IGT groups supported blunted glucose- and/or insulin-stimulated group. Using pathway models, offer evidence supporting important role physiologic response challenge conclude are reflected perturbation experiments. Given involvement transporters human disease, may contribute improved classification via activities.