| Host-pathogen interactions, especially those involving RNA viruses and bacteria, are often characterized by a convergence of ecological and evolutionary time scales. This work explores how such convergence affects the diversity of a fast-evolving RNA virus, influenza, in different host populations. The first study evaluates molecular evidence for a theory of H3N2 dynamics in humans. There is support for episodically strong, continuous positive selection on the hemagglutinin protein and that previously described punctuated changes in antigenicity are not driven by the addition of glycosylation sites. The neuraminidase, nucleoprotein, and matrix 2 proteins also show evidence of positive selection. The second study analyzes time series of serologically confirmed cases of H3N2, H1N1, and influenza B in patients in present-day St. Petersburg, Russia, from 1969 to 1991 to determine whether there is cross-immunity between heterologous strains. Results suggest a role for cross-immunity, but further investigation is necessary. Differences in intrinsic growth rates and rates of antigenic evolution might explain age-related patterns in incidence by virus type and subtype. The third study investigates the effects of heterogeneity in hosts' immune responses on the outcome of strain competition. When immunodominance is skewed toward a single epitope, coexistence inevitably results. When multiple epitopes can be immunodominant, coexistence, limit cycling, chaotic dynamics, and competitive exclusion can occur. Increasing the diversity and breadth of host responses increases the range of cyclic, chaotic, and exclusive dynamics. The last study considers how host ecology affects the long term evolution of influenza's host range, assuming a tradeoff in the virus's preference for certain forms of host sialic acid receptor. A common outcome is the coexistence of specialists, and this outcome is more sensitive to interspecific transmission rates and host population densities than the strength of the tradeoff. Finally, I map three areas of future inquiry: the ability of spatial dynamics and constant antigenic evolution alone to restrict influenza virus diversity, implications of antibody affinity versus neutralization ability for vaccine development, and long-term strategies to manage influenza virus evolution. These studies show that a phylodynamic perspective will be invaluable in developing better predictive models of influenza. |
