Math Biology Seminar: Caroline Colijn

While vaccines are available and are effective in protecting against colonisation and disease with Streptococcus pneumoniae, their effectiveness is limited by strain (serotype) replacement following widespread vaccination. Understanding the post-vaccination balance of serotypes would present the opportunity to achieve a final population composed of the most benign (non-invasive) strains. However, the complex ecology of the pneumococcus makes it difficult to predict the post-vaccination balance of strains. Recently, Corander et al proposed that there is widespread apparent negative frequency-dependent selection (NFDS) in the pneumococcus (Corander et al 2017 Nat. Ecol. Evol.). Here, we use this principle to develop a deterministic model of pneumococcal strain dynamics, and use the model to make predictions about the ecological response of the pneumococcal population to new candidate vaccine strategies. We find that we can identify formulations that out-perform existing formulations in the model. Furthermore, it is possible to obtain a final model population that scores as well as the currently used formulation, using a vaccine strategy with fewer serotypes -- these formulations would be much less costly to produce than current vaccines. We suggest that this approach could provide a template for principled vaccine design based on global surveillance data and genomics. This is joint work with N. Croucher.