Fig. 9

Co-evolution of the C1/C5/ecto-gp41 domains in HIV-1 subtype A and D isolates. Scatter plots demonstrating the correlation between the pairwise distance matrices of viral isolates were built from multiple sequence alignments of unique subtype A1 and subtype D env sequences (n = 10,296, (a) and (b); n = 4657 (c) and (d), respectively). Pairwise comparisons were performed with all pairs of conserved C1, C2, C3, C4, and C5-gp41 regions but only the comparison of C1 + C5/gp41 (a and c) and of C1 + C3 (b and d) are presented in this figure. Pearson product moment correlations are shown for each panel but due to high number of sample sequences, p values were highly significant (<0.0001) for all paired analyses suggesting some level of co-evolution throughout the protein as expected. The heatmaps in (e) and (f) reflect the tree correlation coefficients derived from a similarity comparison of the maximum likelihood derived trees of the multiple sequence alignments of the subtype specific conserved regions. The correlation coefficients between C1 and C5-gp41 demonstrate the strongest similarity of tree topology among the regions compared