Funded Grants

Transient pathogen evolution in heterogeneous host populations

Background

The pathogens that cause infectious diseases, whether newly emerging or long-standing, demonstrate huge potential to evolve. We are left with a long list of questions: Why, every year, do we need to update the human influenza vaccine? Why was the recent introduction of West Nile virus to North America associated with increased virulence in birds and humans? And how has the H5N1 ‘bird flu’ virus been able to show an increase in pathogenicity in ducks over just a few years? Pathogen evolution is happening on a short time-scale, with global consequences for human and animal health.

Pathogens demonstrate this rapid evolution due to a combination of high mutation rates (e.g. imperfect replication of RNA viruses causes mutations to occur very frequently), short generation times (relative to host lifespan) and strong selection pressure (there is a high cost to failing to infect, and multiple ways in which failure can occur). The first two of these processes are responsible for generating multi-strain pathogen populations and the third process regulates that population.

Dealing with multi-strain pathogens is, naturally, more complicated than a single-strain organism. However, research efforts have made great progress in embracing the complexity and describing the dynamics of multi-strain pathogens. Explaining how these pathogen populations will evolve over time, however, has remained a difficult intellectual challenge.

Although such evolution may seem unpredictable, it occurs in a complex system of interacting ecological, population, epidemiological and evolutionary processes. When viewed as a set of component mechanisms acting to change the composition of pathogen and host communities, we can begin to understand how evolutionary forces will act. Host populations are hierarchical and may even be composed of multiple species. Within a species, the population size and the composition of the population are important drivers of host-pathogen interactions. As well as natural selection and mutation, by including consideration of host populations as resources (of variable size and quality) for pathogens, we gain insight into the factors that shape pathogen communities.

Most of the theoretical research in evolutionary epidemiology has focused on understanding the long-term trends of pathogen populations. These models deal exclusively with situations in which a pathogen has established a stable endemic equilibrium in a host population and then consider whether a novel strain arising through mutation is able to displace the resident strain. Whilst these techniques remain valuable guiding tools, we now have the opportunity to develop these ideas to consider more realistic, complex scenarios in which the epidemiological and evolutionary dynamics occur on a similar time-scale and interact with each other.

Mathematical and computational methods provide a quantitative framework for expressing the change in frequencies of pathogen strains in terms of fitness; relating fitness to specific pathogen traits; and decomposing ‘evolution’ into component forces (specifically natural selection, mutation and environmental change) Using these techniques we can understand the dominant processes at play in shaping the evolutionary dynamics of pathogens and begin to make predictions about how they will evolve in a variety of real-world settings. Against the backdrop of emerging pathogens and their demonstrable ability to evolve to more ‘problematic’ forms, this is clearly the time to take up the challenge.

Opportunities for advancement

This project will bridge the gap between traditional theoretical methods for considering pathogen evolution and the emerging data on rapid disease evolution. It will begin by developing a conceptual framework to predict the trait dynamics of multi-strain pathogen populations in realistic situations. Pathogen traits include transmissibility, infectious period and virulence. The keystone of this framework is the partitioning of forces that change mean pathogen fitness into the components of natural selection, mutation, and environmental effects and establishing the relationship between fitness and trait values. Further, it complements traditional epidemiological approaches, which are centered on the per generation measure of pathogen fitness (the basic reproductive number, R0) by instead measuring pathogen fitness per unit time, which enables a joint examination of epidemiological and evolutionary dynamics.

Following the development of this framework, I will consider the application of this method to four central problems in evolutionary epidemiology:

  • How do pathogens evolve when the host population is composed of different types of individual (e.g. vaccinated, naïve, previously infected during childhood)?
  • Do we cause pathogen evolution when we intervene with treatments? How can we balance the benefits of protecting the current population against the costs of selecting for pathogen strains that resist treatment or become more harmful?
  • Do pathogens evolve differently during a cross-species event, where they acquire a new host species with different ecology, population dynamics and immune status?
  • How are predictive models of evolutionary epidemiology compromised by the incomplete information that we are likely to obtain from data? And what can we do to maximize their utility to the scientific community and the general public?

As the project answers this ensemble of linked research questions, it will advance the nascent field of evolutionary epidemiology whilst at the same time providing tools that contribute to the global challenge of controlling infectious diseases. Among the benefits of this project will be the development of techniques to address a broad range of infectious diseases. However, it is motivated by real epidemiological issues and diseases. Indeed, unique equine influenza data have helped shape the project, as, for the first time, they allow us to relate antigenic changes in influenza strains to the:

  • Probability that a host becomes infected
  • Probability that an infected host can transmit the infection
  • Duration of the infectious period
  • I will use this, along with other available data sources, to apply the research output to specific infectious diseases, each with their own key issues.

Broader impact

The techniques that will be developed are extensible to a range of complex systems in which there is (co)evolution, fitness-determining environmental fluctuations, and species interactions. These include the fields of biological invasions, predator-prey systems, food-webs and community ecology.

Here, I focus on a pathogen’s host resource as the key environmental dynamic, but the methods would apply equally well to other biotic organisms or abiotic factors such as climate, opening up a wide range of topics for future study.

Components of this project will also contribute to the general problem of dealing with uncertainty in complex biological systems. As mathematical and computational techniques continue to offer insight into the natural world it is important to also consider how they can best be used against a backdrop of environmental and population-level variation, as well as incomplete and imperfect information that we are able to collect by sampling natural systems.