Figure 3

The effect of eliminating each agent type from the simulated immune response at initialization. Figures 3a, 3b and 3c show the percent of simulation runs that ended with the immune win, loss and hyper-response outcomes, respectively, when the indicated agent type was missing, in combination with initial conditions of 20, 50 or 80 DCs. The control has all cell types present. The number of simulation runs for each data bar is as follows: No Bs with 20 DCs, n = 82; with 50 DCs, n = 73; with 80 DCs, n = 92; no CTLs with 20 DCs, n = 64; with 50 DCs, n = 61; with 80 DCs, n = 106; no DCs, n = 100; no MΦs with 20 DCs, n = 50; with 50 DCs, n = 55; with 80 DCs, n = 76; no NKs with 20 DCs, n = 53; with 50 DCs, n = 71; with 80 DCs, n = 66; no Ts with 20 DCs, n = 75; with 50 DCs, n = 50; with 80 DCs, n = 50; no Granulocyte agents with 20 DCs, n = 93; with 50 DCs, n = 54; with 80 DCs, n = 54. The asterisks indicate significant differences from the control conditions using the Chi-squared test. The p-value for the bars marked **** is p <= 0.0001.