Discrimination by optimal initial conditions. A] Outputs of the simulations of the models with wild type (continuous lines) or levels of CheY4 (total) increased to 5-times normal levels at the start of the simulation (dotted lines). The coloured lines represent the models defined in Figure 1. B] Experimental outputs from tethered cell experiments, wild type output (continuous line) and CheY4 over-expressed in the cells at an average of 5-times the concentration found in wild type cells (dotted line). C] Robust invalidation of the blue model. To ensure that our invalidation was robust we ran 500 simulations assuming that CheY4 is 5-fold over-expressed. We vary parameters k1 - k14 by ± 15% and still observe the same clear difference between the behaviour of the blue and of the red model.