Linking alterations in metabolic fluxes with shifts in metabolite levels by means of kinetic modeling

The links between metabolic dysfunctions and various diseases or pathological conditions are being increasingly revealed. This revival of interest in cellular metabolism has pushed forward new experimental technologies enabling the characterization of metabolic phenotypes. Unfortunately, while large datasets are being collected, which encompass the concentration of many metabolites of a system under different conditions, these datasets remain largely obscure. In fact, in spite of the efforts to interpret alterations in metabolic concentrations, it is difficult to correctly ascribe them to the corresponding variations in metabolic fluxes (i.e. the rate of turnover of molecules through metabolic pathways) and thus to the up- or down-regulation of given pathways. As a first step towards a systematic procedure to connect alterations in metabolic fluxes with shifts in metabolites, we propose to exploit a Montecarlo approach to look for correlations between the variations in fluxes and in metabolites, observed when simulating the response of a metabolic network to a given perturbation. As a proof of principle, we investigate the dynamics of a simplified ODE model of yeast metabolism under different glucose abundances. We show that, although some linear correlations between shifts in metabolites and fluxes exist, those relationships are far from obvious. In particular, metabolite levels can show a low correlation with changes in the fluxes of the reactions that directly involve them, while exhibiting a strong connection with alterations in fluxes that are far apart in the network.