Using disease models and viral sequence data to estimate influenza superspreading

Transmission heterogeneities, arising from variation in contact rates and in host infectivity, are known to play an important role in the ecological and evolutionary dynamics of many infectious diseases. For example, studies have shown that variation in the nature and in the number of contacts between hosts can affect a pathogen's probability of successful invasion into naïve host populations and how quickly establishment occurs. Once a disease is endemic, transmission heterogeneities will also impact the pathogen's evolutionary dynamics, including virulence evolution and the evolution of antigenic diversity. Quantifying transmission heterogeneity is also critically important for guiding control strategies, especially if individuals that are anticipated to spread most of the disease can be identified. Despite its importance, the extent of transmission heterogeneity for any particular disease is difficult to quantify. This is especially the case for diseases with asymptomatic infections and for diseases that are already endemic. Furthermore, the data currently used to estimate transmission heterogeneity are often expensive to collect. Here we show how viral sequence data, in conjunction with epidemiological data, can be used to accurately estimate the degree of transmission heterogeneity in rapidly evolving RNA viral diseases. We apply our statistical approach to seasonal influenza (subtype A/H3N2), demonstrating that individual host variation in this virus's transmission is large. Finally, we discuss possibilities for using our results to identify the superspreaders of influenza.