Pathogen evolution under natural selection: The influenza A case study

Abstract

Recent advances in molecular biology and medicine have provided us with a previously unavailable opportunity to study large amounts of genetic data from various organisms. Sequences of microbial pathogens are subject to a particular interest because the analysis of this data can help achieve a twofold goal. First, insights gained from such analysis help us mitigate the impact of the rapidly evolving pathogens on people’s health. Second, microbial pathogens evolve at rates much higher than those typical for mammals, which allows us to observe evolution in real time and better understand the evolutionary processes in general. My dissertation contributes to the understanding of the role of natural selection in the evolution of rapidly evolving pathogens, in particular of the influenza A virus.

In Chapter 2 I develop a statistical framework for studying the fitness landscape of an organism based on genetic sequence data. The analysis of the influenza A hemagglutinin sequences reveals that positive selection occurs with a strong preference with regard to the target amino acid.

In Chapter 3 I analyze the substitution processes that lead to the formation of amino acid clusters. I show that sequence cluster patterns strongly depend on the underlying phylogenetic relationship and are practically independent of the details of the substitution process.

In Chapter 4 I investigate the validity of the assumption of neutrality of synonymous substitutions in the influenza A genome. Using the fact that the topology of phylogenetic tree has information about selection pressures acting on the organism, I show that the synonymous nucleotide composition of influenza A has been changing in a way that cannot be explained without invoking natural selection.

In Chapter 5 I develop a theoretical framework for modeling epidemiology and evolution of multi-strain pathogens. The suggested approach allows us to construct tractable multi-strain pathogen models under a wide variety of assumptions. Using this approach, I suggest a simple mechanism of how frequency-dependent selection shapes the evolution of a virus with two epitopes that elicit independent immune responses.