CTL dynamics and the impact of recombination.

<p>Figures are plotted as a function of mean avidity and variance of avidity of the CTL response. Panel A shows the mean diversity of the CTL population which decreases with increasing mean (except for an initial increase in avidity) and decreasing variance in avidity. Panel B shows the mean number of escapes during infection without recombination after 1 yr. of infection, which has an optimum for a given mean avidity of response and correlates well with the diversity of the CTL response in panel A. Panel C shows the factor of increase in the number of escapes after one year (cf. panel B) due to recombination, which is the strongest for broad and even responses. Panels D and E show the time of appearance (days) of the full escape variant without and with recombination, respectively. The hatched areas show runs in which the full escape mutant did not appear within the time of 1000 days. The impact of recombination on the time of emergence of the full escape mutant, shown in panel F, corresponds well to the impact of recombination on the number of escape mutations after one year, shown in panel C. The mean and variance of log avidities ( and , respectively) plotted in all panels are defined in each simulation, and the distribution of avidity values is drawn from a uniform distribution in each simulation run between and . The parameters used: , , . All other parameters are given in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0016052#pone-0016052-t001" target="_blank">Table 1</a>. Each point in the plot shows an average of 5000 simulation runs.