Molecular epidemiology and evolution of the 2009 H1N1 influenza A pandemic virus
The swine-origin H1N1 influenza A pandemic virus (A(H1N1)pdm09) was detected in the human population in March 2009. Due to its antigenic novelty, the majority of individuals were susceptible to the virus and the pandemic quickly disseminated around the globe. Rapid characterization of the epidemic was required in order to help inform interventions and determine the risk posed to public health. Widespread sampling and sequencing of virus isolates enabled early characterization of the virus using phylogenetic analysis and continued surveillance over the subsequent three years of global circulation. Throughout this thesis, Bayesian phylogenetic methods are employed to investigate how quickly evolutionary parameters can be accurately and precisely estimated from pandemic genome sequence data and explore how selection has acted across the A(H1N1)pdm09 genome over its period of transition to a seasonal influenza lineage. It is shown that accurate estimates of the evolutionary rate, date of emergence and initial exponential growth rate of the virus can be obtained with high precision from analysis of 100 genome sequences, thereby helping to characterize the virus just 2 months after the first cases were reported. In order to account for variation in growth rates of influenza epidemics between localized outbreaks around the globe, a hierarchical phylogenetic model is employed for analysis of pandemic and seasonal influenza data. The results suggest that the A(H1N1)pdm09 lineage spread more easily and with greater variation between populations during its first pandemic wave than either seasonal influenza lineage in previous seasons. The birth-death epidemiology model has been shown to provide more precise estimates of the basic reproductive number than the coalescent in analysis of HIV epidemic data. Analysis of pandemic influenza data carried out here suggests that the model assumptions are less applicable to influenza and in fact thebirth-death epidemiology model loses accuracy more rapidly than coalescent models as data increased during the pandemic. The effects of an increasingly immune global host population over the pandemic and subsequent influenza seasons were investigated using robust counting of substitutions across the genome. Results suggest that antigenic genes were under a greater selective pressure to evolve than internally expressed genes and the rate of non-synonymous substitution was highest across all segments immediately after emergence in the human population. Bayesian phylogenetics is increasingly being employed as an important tool for rapid characterization of novel infectious disease epidemics. As such, the work carried out here aims to determine the accuracy and applicability of existing evolutionary models with pandemic sequence data sampled over a range of temporal and spatial scales to help better inform similar analyses of future epidemics.