Fig. 5

Jensen–Shannon divergence identifies positions of differential amino acid usage between R5 and X4 HIV-1 Tat sequences. HIV-1 Tat sequences were sorted into R5 (n = 504) and X4 (n = 31) populations according to the predicted co-receptor usage of the co-linear V3 domain as determined by Web-PSSM score. a The diversity index at order = 1 was calculated for each position for both R5 (red) and X4 (blue) Tat sequence populations. The diversity index between R5 and X4 populations displayed high similarity at nearly all positions, with the second half of Tat displaying higher diversity values overall for both populations. b The Jensen–Shannon divergence between R5 and X4 Tat sequences was computed for each amino acid position and plotted with a diamond. Statistically divergent positions 7, 23, 57, and 60 (P < 0.01) were plotted in red and consensus changes, positions 40 and 67, were plotted in yellow. A Monte Carlo permutation test was performed to iteratively group Tat sequences into random groups and calculate a distribution of expected Jensen–Shannon divergence values. The full range of this distribution was plotted in light blue with the interquartile range plotted in dark blue