Date Published: January 3, 2013
Publisher: Public Library of Science
Author(s): Daniel Zinder, Trevor Bedford, Sunetra Gupta, Mercedes Pascual, Neil Ferguson.
Influenza A (H3N2) offers a well-studied, yet not fully understood, disease in terms of the interactions between pathogen population dynamics, epidemiology and genetics. A major open question is why the virus population is globally dominated by a single and very recently diverged (2–8 years) lineage. Classically, this has been modeled by limiting the generation of new successful antigenic variants, such that only a small subset of progeny acquire the necessary mutations to evade host immunity. An alternative approach was recently suggested by Recker et al. in which a limited number of antigenic variants are continuously generated, but most of these are suppressed by pre-existing host population immunity. Here we develop a framework spanning the regimes described above to explore the impact of rates of mutation and levels of competition on phylodynamic patterns. We find that the evolutionary dynamics of the subtype H3N2 influenza is most easily generated within this framework when it is mutation limited as well as being under strong immune selection at a number of epitope regions of limited diversity.
Influenza viruses are classified into types A–C, among which influenza A is the most pathogenic. These viruses cause between a quarter to half a million deaths worldwide  and tens of thousands of deaths in the US during annual epidemics . The economic burden of seasonal influenza in the US is estimated at more than ten billion dollars in healthcare costs alone .
We use an individual-based SIR model, explicitly tracking the chains of infection of viral lineages as well as the antigenic phenotype of every virus in the population (Figure 2A). Thus, our model explicitly tracks viral genealogy rather than conducting phylogenetic inference, and therefore does not include any genotype to phenotype map.
Herein, we implemented an individual-based model that allowed us to track both the ecological and evolutionary dynamics of a pathogen population, in which cross-immunity is orchestrated by a finite set of antigenic loci of limited variability . We used this model to compare phylodynamic patterns under a regime governed primarily by limitation on the introduction of antigenic mutations (mutation limited), to a regime determined by the availability of antigenic niches (selection limited), and under varying strengths of competition between strains. We use this framework to determine the conditions under which a limited antigenic repertoire could explain the observed phylodynamic patterns of H3N2 influenza.