Fig. 3: Simulations of the PDE model (10) with reduced function fr.
Empirically reduced single cell state variable x, showing that the reduced system captures the full EMT network dynamical response; and the combined influence of biochemical noise with input signal dynamical response in modulating the phenotypic heterogeneity of the cell population. Ai Bifurcation diagram showing stable (blue) and unstable (red) levels of miR-200 (y axis on left) for increasing SNAIL levels; and distribution of SNAIL levels among the cells (y axis on right) for different mean characteristics levels S0. Aii Distribution of cells along the state x of reduced EMT system for different distributions of SNAIL shown in (Ai) after a simulation time of 100 days starting from epithelial cells. B Changes in population’s phenotypic distribution with increasing levels of epigenetic noise ηx for SNAIL distributions with mean S0 = 190K (Bi) and 225K (Bii). For S0 = 175K (Biii), the population remains invariant to increasing noise. C Temporal changes in cell state distribution of the population for decreasing values of input signal SNAIL's perturbation recovery rate (increasing values of the characteristic time ‘α’). D Time evolution of population’s heterogeneity (measured by differential entropy) for different α parameter values. Parameters used to generate plots, unless stated otherwise, are α = 120 h, ηx = 1000, ini pop Epi (see Fig. 4A, i), time = 100 days, S0 = 225K molecules, and division rate r is constant across phenotypes, given by r = 0.0182/h.
