Chronic infections can shape epidemic exposure: Pathogen co-occurrence networks in the Serengeti lions

Nicholas M. Fountain-Jones; Craig Packer; Maude Jacquot; F. Guillaume Blanchet; Karen Terio; Meggan E. Craft

We categorized each pathogen as chronic or acute based on the literature and, as the biology of most of these pathogens in lions are poorly understood, age-prevalence relationships. Pathogens which peak in prevalence at a young age (<= 2 y.o.) with little temporal fluctuation were considered to be chronic, whereas an increasing age-prevalence relationship and high temporal variation were classified as more acute (Fig. S1 ). Feline coronavirus can have both acute and chronic forms and it is impossible to assess which form the individual was exposed to from serological data, but based on age-prevalence relationships we categorized it as a chronic infection (Fig. S1 ). As most individual lions were likely to be infected by chronic infections within the first two years after birth (Troyer et al. 2011, Fig . S1), we assume that acute exposure typically occurred whilst the individual was infected by the chronic pathogen. We then partitioned the chronic pathogen data into two sets based on taxonomic resolution (high and medium). The high taxonomic resolution dataset encompassed FIV Ple genotype and Babesia species data, whereas the medium resolution dataset aggregated FIV Ple subtype information and Babesia data to genus level. We created a summary co-occurrence matrix (n*n) (Griffiths et al. 2014) that described the amount of co-occurrence of both chronic and acute pathogens within individuals by multiplying the incidence matrix (i.e., contingency table that described the occurrence of pathogens across individuals) by its transpose. We then computed a measure ( ) of network structure and modularity index based on node overlap and segregation (Strona & Veech 2015) .