The Distribution of Mean Divergence and Maximum and Minimum Estimates of Divergence for Pairwise Comparisons of Alleles

<div><p>Frequency in 1,000 trials plotted against percentage divergence.</p><p>(A) All alleles (representing MTBC versus smooth).</p><p>(B) Recombinant alleles only (representing smooth).</p><p>Percentage divergence is the equivalent of percentage divergence at synonymous sites. The simulation started with an ancestral sequence of 800 base pairs (bp), representing the synonymous sites in the 3.4 Kb of sequence presented by Gutierrez et al. [<a href="http://www.plospathogens.org/article/info:doi/10.1371/journal.ppat.0020098#ppat-0020098-b001" target="_blank">1</a>]. This was used to generate two independent populations with eight derived sequences following three bouts of binary replication. At each replication, a randomly selected bp was chosen and, to simulate mutations, changed with probability of 0.5. This generated, on average, 15 polymorphic sites in the entire dataset of 16 derived sequences, which approximates the number of polymorphic sites in the clonal frame of the dataset presented by Gutierrez et al. [<a href="http://www.plospathogens.org/article/info:doi/10.1371/journal.ppat.0020098#ppat-0020098-b001" target="_blank">1</a>]. To compare differences arising from recombination, only one of the two independent populations (described above) was allowed to undergo recombination. To simulate this, sets of bp were changed in proportion to the observed recombinant segments shown boxed in Figure 3 of Gutierrez et al. [<a href="http://www.plospathogens.org/article/info:doi/10.1371/journal.ppat.0020098#ppat-0020098-b001" target="_blank">1</a>]. For example, in four randomly selected sequences, the same 5 bp were simultaneously changed to simulate the recombinant segment in <i>katG</i>. Recombinant segments did not overlap. The above simulation was run for 1,000 independent trials; starting from the same ancestral sequence and the average pairwise diversity, the maximum and the minimum pairwise diversity were recorded for each population, for each trial.</p>