Modeling Virus Transmission and Evolution in Mixed Communities

Abstract

In the early 1980s Anderson and May showed that parasite virulence (host mortality rate when infected) and parasite transmission are positively correlated because of their joint dependence on host exploitation (e.g. replication rate). This correlation often results in maximum parasite fitness at intermediate virulence, which has important implications for both parasite evolution and transmission. Anderson and May's observation has led to nearly four decades of work on the ecology and evolution of host-parasite interactions, which focuses on making either general predictions for a range of simplified host-parasite systems or detailed predictions for a single host-parasite system. Yet, despite decades of research, we know comparatively little about parasite evolution and transmission in heterogeneous and/or small host populations. Additionally, much previous work has distanced itself from empirical data, either by outpacing the collection of data or under-utilizing available data. My work focuses on the evolution and transmission of parasites in heterogeneous host populations; I rely on tradeoff theory, but adopt a case-study approach to maximize the use of empirical data. Using West Nile virus infections of birds I show that a continent-wide strain displacement event cannot be explained by current data (Chapter 2), and that transmission in heterogeneous host communities can be estimated using data from citizen scientists, laboratory experiments, and phylogenetic comparative analysis (Chapter 3). Using Myxoma virus infection of European rabbits, I show that tradeoff theory can help us to understand parasite evolution in host populations with heterogeneous secondary infection burden (Chapter 4). In Chapter 5 I show that poorly evolved parasites invading new host populations experience transient evolution away from optimal virulence. In addition to my biological focus, I emphasize clarity and rigor in statistical analyses, including the importance of appropriate uncertainty propagation, as well as reproducible science.