A cloud theory for metabolic regulation. The PSEUDO method hypothesizes the existence of a degenerate optimal region. Fluxes are regulated to approach this region, but allowed to vary freely within it. (A) Variability of fluxes within the degenerate optimal region. As described in the methods, we used Markov chain Monte Carlo sampling to produce 3000 randomly distributed fluxes, each consistent with at least 99% maximal growth. 3 specific fluxes are plotted on these axes, colored to emphasize their value in each dimension. Red, green and blue content corresponds to value in x, y and z respectively. Histograms present overall distributions of individual fluxes. (B) Variation in measured values for 24 fluxes across 91 transcription factor deletion mutants from the Sauer data set. Measured data is plotted similarly to the Monte Carlo generated data. (C, D) Computationally estimated flux variability within the degenerate optimal region correlates well with measured variability in both the Sauer (C) and Tomita (D) data sets. The coefficient of variation (CV) of the distributions is a normalized expression of both measured and predicted flux variability. Marker shapes are used to indicate the metabolic class of each flux. Metabolic fluxes that are only weakly coupled to biomass production can vary widely within the degenerate optimal region. The same fluxes are more likely to vary in under perturbation in published data sets.