Multiscale modeling of influenza viral emergence

Influenza is viral disease characterized by a sudden onset of fever, cough, headache, muscle and joint pain, severe malaise, sore throat and a runny nose. While most people recover from these symptoms in about a week without medical attention, there are cases with complications where the illness is severe and can cause death. Influenza spreads around the world in yearly outbreaks, causing about 3 to 5 million cases of severe illness and about 290,000 to 650,000 deaths due to respiratory complications (from WHO). The World Health organization suggests the use of the vaccine in order to protect the people at higher risk (very young children and elderly people, for example), but the immunity given by the vaccine is not lifelong and vaccinations of one year could be ineffective even the next year. This is due to highly mutating nature of the influenza virus, which generates a huge immunological variability. This means that, following a mutation, the virus can change the antigens that it presents and thus be recognized as of a different “type” by the immune system with respect to its previous configuration of antigens. This effectively creates a new viral strain which can easily infect individuals that were protected to the previous strain. This phenomenon, that is called immune escape, is detrimental to the production of the vaccine, which requires the identification of the dominant strain on each year. An error in the prediction of the dominant strain therefore might occur, and cause higher number of reported cases during the season. This thesis tries to address this problem by developing a model of the spreading of two strains of the influenza virus in a population of hosts. This process is based on the interplay of within-host factors and the between-host transmission, and both scales are accounted in the model. Because of the resulting complexity of the model, a numerical approach has been devised in order to study the outcomes of the epidemics, averaging the results over many numerical simulations. This required transforming the within-host model, originally deterministic, into a stochastic simulation, and implementing a set of rules for the transmission of the two strains between individual hosts. The model has been studied first with only one viral strain, exploring the space of the parameters of the model, both within-host and between-hosts. This first stage was required in order to understand the results and reproduce realistic outcomes of influenza epidemics. Afterwards, the second strain was introduced in the simulations, and the effect of parameters of this emerging strain was studied by exploring their possible values. As a main result, it was found that the frequency on which the hosts contacted each other played an important part in the selection of the dominant strain.