Objective: To estimate the relative substitution rate of the individual positions in an alignment of HIV-1 env sequences coding for areas V3, V4, V5, and the beginning of gp41, and to study phylogenetic relationships between HIV-1 strains taking into account these substitution rate estimates.
Design: Phylogenetic comparison of 145 HIV-1 strains classified in HIV-1 group M, subtypes A-H and isolated from patients of 24 different geographical origins.
Methods: A new method recently developed for measuring the substitution rates of the individual nucleotides in a sequence alignment was applied to an alignment of env gene sequences. From the resulting substitution rate distribution, an equation was derived that describes the relationship between dissimilarity and evolutionary distance better than equations previously available. Phylogenetic trees were then constructed from matrices of distances computed using this new equation.
Results: 'Substitution rate calibration' offers detailed information on the relative substitution rate or variability of the nucleotides in the env gene. A large phylogenetic tree of 145 env gene sequences constructed by neighbour-joining and taking into account the substitution rate spectrum for this gene, clearly shows the existence of the eight subtypes A-H, all supported at a bootstrap level of 90% or higher. Intersubtype distances were between 0.25 and 0.38, which is considerably higher than those found in trees not considering differences in substitution rates among different alignment positions.
Conclusions: Evolutionary distances are seriously underestimated when individual substitution rates are not considered in the estimation evolutionary distances. Furthermore, due to the more accurate estimation of evolutionary distances, naturally occurring HIV-1 intersubtype recombinants could be recognized more easily.