Molecular phylogenetics of clonal and non-clonal populations of bacterial pathogens.

摘要

Estimations of evolutionary relationships among samples rely on the identification of characters that are shared among some, but not all samples. For molecular characters, these loci must have mutated during the time span of evolution under analysis. If this time span is large, then there is a potential for subsequent mutations to occur, confusing efforts to reconstruct phylogenies. Conversely, a short time span may not offer enough time for mutations to arise in the first place. It is therefore imperative that the mutation rates of molecular loci be taken into account as faster mutating loci can be used to determine relationships among closely related samples while slowly mutating loci are best used for resolving ancient relationships.;I therefore used variable number tandem repeat (VNTR) loci for reconstructing the phylogeny of groups of Burkholderia pseudomallei samples collected from single infections or outbreaks in Australia. Not only were these loci able to discriminate among very closely related samples such as within a single patient, but they were also useful for determining how such isolates were related to each other. However, these loci should not be used for determining how major groups are related to each other as subsequent mutations make it more likely that unrelated samples will share allelic states by chance alone. Furthermore, B. pseudomallei cells can exchange DNA and incorporate exogenous DNA into their genomes. This lateral gene transfer also confounds efforts to determine deep levels of relatedness.;In order to reconstruct ancient relationships despite the probability that different portions of the B. pseudomallei genome may have different evolutionary histories through lateral gene transfer, I used single nucleotide polymorphisms (SNPs) as phylogenetic characters. SNPs have a relatively low mutation rate and are thus evolutionarily stable; however they are also relatively rare in the genome and therefore require extraordinary efforts such as whole genome sequencing to find. By using tens of thousands of SNPs scattered around the genome, the phylogenetic effects of separate evolutionary histories of individual genes were overwhelmed by the evolutionary history of the majority of the genome. I therefore used the whole genome sequences of 43 isolates to determine deep evolutionary patterns of B. pseudomallei and its close relatives.;The comparison of whole genome sequences for phylogenetic inference represents the ultimate tool for discovering relationships. However, sequencing an entire genome remains expensive and time consuming, making it more beneficial to integrate whole genome genotyping with sub-genome genotyping. Such methods result in highly accurate phylogenies whose topologies are subject to phylogenetic discovery bias. I therefore explore this bias using a simple model of clonally propagating bacteria that are not subject to lateral gene transfer. In doing this, I explore ways to further increase the accuracy of phylogenies for strictly clonal populations.