Figure 6.

The effect of assuming the wrong population parameters on the escape rate estimates. To quantify the robustness against wrong assumptions, we simulate escape dynamics with parameters different from those assumed in the escape rate estimation. (A–C) show simulations with N = 10⁷ and μ = 10⁻⁵ per day, while (D–F) use a 10-fold higher mutation rate μ = 10⁻⁴. In (C,F), the simulated recombination rate varies as shown. (A,D) Assuming a too small population size results in estimates that are too large. The effect is more pronounced at lower mutations rates. (B) Similarly, if the mutation rate is assumed too large, the estimated seeding of multiple mutants occurs too early and the estimates of escape rates are too low. Note that assuming the correct rates [μ = 10⁻⁵ in (B) and μ = 10⁻⁴ in (F)] results in unbiased estimates. (C,F) If the population recombines, the actual seed times are smaller than those estimated by the fitting routine. To compensate for the shorter time interval during which the escape variant rises, the estimates of escape rates are larger than the actual escape rates, at least at low mutation rates. For high mutation rates, recombination is less important because additional mutations are more efficient at producing multiple mutants than recombination. Mean and standard deviation at each point are calculated from 100 independent simulations.