Fig. 2.

a) The probability that a WT reversion mutation occurs before the compensatory mutations (Com) fix in the population. In these simulations, clones with compensatory mutations cannot acquire a second mutation (i.e. they cannot revert to WT). Simulations are run for different values of μ, the mutation rate, and different values of p, the compensatory effect. Simulations are run for 500 generations and for each parameter combination we run 1,000 simulations. Lines are based on the mathematical model. b) Same as (a), but simulations are run for different values of n, the target size for compensatory mutations. Simulations are run for 500 generations. Other parameters: population size N = 10,000, cost c = 0.15, effect of compensatory mutations P = 0.5. The WT is considered to be present (and expected to fix) if at least present in 10 individuals at the end of the simulation (0.1% of the population).