Assessing the phylodynamics of RNA viruses which cause acute disease
Open Access
- Author:
- Stack, Joseph Conrad
- Graduate Program:
- Biology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- None
- Committee Members:
- Ottar N Bjornstad, Committee Chair/Co-Chair
Edward C Holmes, Dissertation Advisor/Co-Advisor
Beth Shapiro, Dissertation Advisor/Co-Advisor
Matthew Joseph Ferrari, Committee Member
David Russell Hunter, Committee Member - Keywords:
- evolution
disease
infection
epidemiology
phylodynamics
coalescent - Abstract:
- Phylodynamic methods, applied to viral genetic sequence data, have been utilized with tremendous success in understanding aspects of the epidemic histories of many RNA viruses. They have, however, been applied almost exclusively at the epidemiological scale and there exist numerous challenges to be addressed in order for them to be successfully applied at lower spatiotemporal levels. These challenges arise from (1) the lack of understanding about what information genetic sequences contain regarding RNA virus epidemic dynamics, which are often complex and not well-understood from an evolutionary standpoint, and (2) the fact that viral evolution manifests in the viral genetic sequence data much differently at higher levels than at lower spatiotemporal scales. In regard to the first of these challenges, I explore how the protocol used to sample viral genetic sequences affects the quality of phylodynamic inference. Using large-scale epidemic and evolutionary simulations, I find that viral genetic data sets collected under different protocols can return very different inferences of a virus's epidemic history. Based on these findings, I suggest how data might be collected so as to maximize their utility in phylodynamic analyses. A prerequisite for adapting phylodynamic methods to lower spatiotemporal scales is having a high-resolution picture of epidemic dynamics at these scales and some understanding about how these dynamics correlate with observed patterns of viral evolution. I present two research studies to help address this need. I present an extensible, Bayesian method which employs minimal epidemiological data to make inferences about population structure by selecting between different contact network models. This method is rigorously evaluated using both simulated and empirical data and I find that it is helpful in determining aspects of host population structure, given reasonable assumptions. I also evaluate how viral genetic sequence data correlate with a highly resolved picture of viral transmission (epidemic) dynamics. Using influenza A virus genetic sequences from three case studies, I demonstrate how certain genetic variants, but not others, accurately correlate with the virus's known transmission history. These findings provide the basis for further phylodynamic analyses and highlight the predominant evolutionary forces that will need to be accounted for.