Results of a typical simulations.

<p><b>A</b>: Results of a typical run at the exponential growth phase and at equilibrium (type A cells are red and type B cells are green). Left panels: “intrinsic” phenotypic switch; right panels: “extrinsic” phenotypic switch.. In all simulations N_{max death} = 40 was used. <b>B</b>: The distribution of the number of neighbours around the A and B cells (upper and lower panels, respectively) in the “extrinsic” (left) and “intrinsic” (right) models. The average number of neighbours and the standard deviation is indicated for each panel. <b>C</b>: Analysis of the spatial randomness of cell distribution at equilibrium using Ripley's L statistics. Ripley's L a point pattern with those generated by a homogeneous Poisson process. The plot shows the estimate of L(h) for various values of R ( = radius of a circle around the cell), and compares them to the line y = 0 (expected in a homogeneous Poisson process). An envelope defining the confidence interval is obtained from the maximum and minimum L(h) estimates of a large number of Monte Carlo simulations. The point set is significantly clustered in the range of scales where the estimated L(h) values are larger than 0 and lay outside the region defined by the envelope. This is the case for type A cells when considering small distances in the “extrinsic” model (red line). However, there is no significant clustering of type A cells (red line) in the “intrinsic” model and no clustering of type B cells (green line) in any of the two models, because all L(h) values are close to 0 and lay inside the region defined by the envelope (black lines).