Figure 1

Brief illustration of our main results. (Top layer) the predictions of stability analyses based on dynamical systems theory—systems should tend to converge toward SSS, as shown around SSS points by simulation results of Langevin equations, while keeping away from USS, but the realities are contrary to this prediction. (Middle layer) the fluorescence detection results in biochemical experiments; this real system has an additional stable state nearby the USS. (Bottom layer) the probability density functions of exact stochastic simulations’ results; they are consistent with experimental results.Reuse & Permissions

Figure 2

Schematic illustration of the genetic toggle switch plasmid. Each promoter is repressed by a dimer of corresponding repressors. GFP and RFP are fluorescent proteins coexpressed with respective repressors, which act as reporters for presenting the current state of this toggle switch.Reuse & Permissions

Figure 3

Key results of FACS experiments. (a) Four groups of differentiation ratios under different conditions. The final concentrations of inducers are listed under every group of bars. In each group, the gray bar represents the population percentage of bacteria that is in the no-fluorescence state; the red bar represents the percentage in the red state, and the green bar for the green state. (b),(c) FACS dot plots in conditions B and C marked in partition (a). Regions Q1, Q3, Q4 correspond to green-, no-, red-fluorescence states, respectively. (d) The restabilized differentiation ratios after sorting. The source regions are listed under every group of bars, from which the bacteria are sorted separately.Reuse & Permissions

Figure 4

Deterministic nullclines (solid curves) of species $A$ and $B$, and their stochastic counterparts (dashed curves). Each stochastic one is generated by a series of mean values from stationary conditional distributions, which are shown in the inset histograms. (a),(b) without induction and (c),(d) with appropriate induction.Reuse & Permissions