Figure 2

Empirical phylogenetic allelic structure, simulated parsimony score and scaled MI. (A) Example, ML phylogenetic trees of AMA-1 with branches coloured on the basis of observed combination of amino acid variants found at tips for different pairs of sites. Pink and cyan are used for the two most common amino acid variants (which are discordant for these sites, see Table 1) and black for the least common. (B) Empirical pairs of sites described in Fig. 1 and Table 1 are shown in relation to a null distribution of site-pairs from 1000 simulations that scored most highly (rank 1) on S-score (the product of scaled MI measuring non-overlap, and geometric mean of parsimony score indicating the minimum number of genetic changes required to explain the observed ancestral relationships, GMPS). Color key (scale) on the right is the two-dimensional density of GMPS and scaled MI of the simulated site-pairs. (A,B) Pairs of sites identified in red and green in the two panels match the pairs in Fig. 1.