The concept of species in bacteria is a matter of contention. The current definition is based on DNA-DNA hybridisation and does not account for evolutionary forces that are important in demarcating species. In this thesis we investigate two evolutionary forces that are important in speciation in bacteria, propose novel statistical models for them and infer parameters of interest. We present the first attempt at inferring the bias in the recombination process from whole bacterial genomes. Despite empirical evidence that recombination is biased and theoretical results that this bias is important in speciation, it is usually ignored. We propose a coalescent based model that accounts for the bias in the recombination process. We use approximate Bayesian computation for inference and describe an efficient method for simulating from the model. We show that our method performs well on simulated datasets and is robust to slight misspecification of the history of the samples. Application of our method to a Bacillus cereus dataset shows that it contain evidence that the recombination process depends on the evolutionary distance between donors and recipients. We demonstrate that the rate of bias in the recombination process for this dataset is far lower than what theoretical studies require for the spontaneous generation of populations that can be called species under neutral model. Next we propose a model for occurrence of adaptive events on a phylogenetic tree. We use the model to infer the boundaries of clusters on a phylogenetic tree that correspond to ecologically distinct lineages. we characterise our method using simulated datasets and show that it is conservative in estimating the number of adaptive events. Finally we apply our method to two bacterial datasets of Salmonella enterica and Vibrionaceae. We show that there is decisive evidence that isolates in these datasets partition into numerous ecologically distinct lineages and use our method to delineate the boundaries of these lineages.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:711762 |
| Date | January 2014 |
| Creators | Ansari, M. Azim |
| Contributors | Didelot, Xavier ; Bowden, Rory |
| Publisher | University of Oxford |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | https://ora.ox.ac.uk/objects/uuid:b830a37a-fa7e-4b68-9868-fc5c629d45f5 |
Page generated in 0.0025 seconds