Origin of HBV and its arrival in the Americas--the importance of natural selection on time estimates.

BACKGROUND The strong geographic structure shown by the global pattern of HBV lineages suggests an ancient origin for this virus; however, estimates based on the molecular clock suggest a very recent origin for the Native American genotypes F and H. In this study, we contribute to this debate by estimating the divergence times of genotypes F and H and by discussing how evolutionary rates estimated from recent samples may underestimate the divergence time of more ancient nodes in HBV phylogenies. METHODS A total of 108 complete HBV genotype F and H genomes were compared to 44 reference genomes from other genotypes. Time estimates were based on a Bayesian method with evolutionary rates taken from the literature. To assess the pattern of substitutions in recent versus old branches we mapped the phylogenetic distribution of all mutations occurring in genotypes F and H using a maximum likelihood approach and compared the number of synonymous and non-synonymous mutations in young and old branches of HBV genotype F and H phylogeny using a χ² test. RESULTS Estimated divergence times between genotypes F and H depend heavily on the evolutionary rate. While fast rates suggest a recent separation of these genotypes (approximately 800 years ago), slow rates suggest an earlier divergence (up to approximately 13,000 years ago). There is a clear excess of non-synonymous substitutions in the most recent branches of HBV phylogeny (P=4.87×10⁻¹⁵), most likely suggesting the action of purifying selection. CONCLUSIONS These results suggest that rates estimated based on recent samples will overestimate the evolutionary rate and underestimate the coalescence times for ancient nodes in HBV phylogeny.