FIGURE 2 

Distribution of metabolic fluxes in the E. coli 's TCA cycle in arabinose minimal medium for (A) wild-type organism predicted by FBA, (B) fbaA mutant predicted by MOMA, (C) optimal state of fbaA mutant predicted by FBA, and (D) fbaA mutant with the rescue deletions of genes aceA and sucAB, predicted by MOMA. Key flux changes are highlighted in orange. Note that the metabolic flux pattern predicted by MOMA after the fbaA deletion (B) is similar to the wild-type fluxes (A). With the rescue deletions, however, MOMA-predicted fluxes (D) are brought closer to the FBA-predicted fluxes (C), restoring the organisms' ability to produce biomass. While we show a double deletion for its pedagogical value, we note that the deletion of aceA alone is sufficient to rescue the mutant (see Figure 3A) and that the mutant can also be rescued with other single-gene deletions (see Figure 4B and Supplementary Information).