In Silico Discovery of Single Nucleotide Polymorphisms Under Positive Parallel Selection in Bacteria

Bacterial pathogen are continuously evolving; most changes are deleterious or neutral, but sometimes a mutation leads to functional change. Such mutations are so very rarely beneficial, that when they do arise in parallel in distantly related isolates (known as homoplasy) it implies that the change is likely positively selected because it confers an adaptive advantage to the bacteria. In this thesis a method is developed to identify and characterise potential homoplasies within a target bacterial population, using large-scale whole-genome data; the method is then applied the method to investigate patterns of positive selection in Mycobacterium tuberculosis.