Uses of statistical parsimony in HIV analyses

摘要

Molecular phylogenies have become powerful tools in human epidemiological studies. Because the phylogeny represents the historical relationship of genes through time, it plays an important role in the elucidation of both historical patterns and processes at work on the gene region of interest, and therefore, on the disease associated with that gene region. However, phylogenetically based analyses are only as good as the phylogenies upon which they are based. Two common problems result from the application of phylogenetic techniques to the population genetic level; 1) lack of resolution due to the short divergence times of a population study, and 2) incorrect inference due to the comparison of nonhomologous sequence regions resulting from recombination. A population based method for reconstructing historical relationships among gene sequences is statistical parsimony. In this paper, I outline the limitations of traditional methods, outline the advantages and demonstrated superiority of statistical parsimony when divergences among sequences are low. Finally, I demonstrate the multiple applications of this estimation procedure to problems relating to human immunodeficiency virus evolution.