Fig. 4
From: Statistical mechanics for metabolic networks during steady state growth
a Predicted fluctuations in the growth flux, σ/λmax, as a function of βλmax. Model inferred from data in Fig. 2a is shown as a blue point and the approximate range experimentally probed in b–d is shown as green shade. In the scaling regime (cyan dashed line), σ ∝ (βλmax)−1. b single-cell measurements (each dot = one division event) of elongation rate in a microfluidic mother machine device for wild-type E. coli under increasing sub-inhibitory antibiotic concentrations, indicated on top; data from a single experiment in ref.37. After switching to higher antibiotic concentration (lighter shade of green), cells show transient behavior (empty green circles) which we ignore and focus only on steady state (full green circles). c Measured distributions (plot symbols) of growth rates from multiple experiments as in b (different shades of green = different tetracycline concentration). Solid lines show maximum entropy distributions with best-fit values for (β, λmax) (Supplementary Note 3). d Testing the predicted scaling of growth rate fluctuations. Each dot = one lineage from experiments in c. The average growth rate, \(\bar \lambda\), the maximal observed growth rate, λmax, and the SD of the growth rates, σ, were estimated separately on each lineage. Scaling regime of a predicts a linear relationship, which is shown here by binned data (red points with errorbars = mean ± SD within equi-distant bins) and best linear fit (solid black line, R2 = 0.77)
