Skip to main content


Figure 5 | BMC Systems Biology

Figure 5

From: An objective function exploiting suboptimal solutions in metabolic networks

Figure 5

Increased TCA cycle predictions using PSEUDO are partially explained by decreased growth, organic secretion and PPP fluxes. (A) Carbon consumed as glucose may ultimately be converted into biomass, oxidized to CO2 or secreted as organic metabolites. For the zwf mutant, FBA predicts the most growth, MOMA predicts the most secretion and PSEUDO the most CO2 production. The molar flux is reported in units of mmol carbon gDW-1 hr-1. (B, C) Sensitivty analysis of growth and organic secretion with respect CO2 output. The WT model achieves optimal growth when CO2 output reaches 15 mmol gDW-1 hr-1. The FBA model of the zwf mutant attains 87% WT growth with a slightly higher optimal CO2 output. The MOMA model predicts a CO2 flux output for the zwf mutant similar to WT, with decreased growth and increased secretion. The PSEUDO objective identifies a wide range of CO2 output fluxes consistent with growth near 80% optimal. The high CO2 output selected by PSEUDO coincides with near-zero carbon secretion, similar to the WT. (D) PSEUDO predicts near-zero flux though the oxidative PPP in the zwf mutant, while both FBA and MOMA predict positive flux. Reducing oxidative PPP flux only marginally decreases growth (<1%) for FBA and MOMA predictions, while significantly increasing MOMA growth predictions. Growth is reported as percent WT. (E) Reducing PPP flux reduces organic secretion in the MOMA model, with no effect on secretion in other models. (F, G) Reduced PPP flux leads to increased glycolysis and TCA cycle fluxes. As the PPP flux approaches zero, the FBA, MOMA and WT predictions converge, suggesting this perturbation moves all three methods to PSEUDO-like solutions.

Back to article page