Reservoir Dogs: Consequences of variable contact network structures for disease spread in free-ranging dogs

The risk of disease spread is contingent on not only the biological properties of causal agents, such as viruses and bacteria, but also the socio-ecological context of the outbreak. Therefore, researchers need to incorporate the variation in animal-animal interactions in models of disease spread. Much research has focused on how ecological factors (seasonality, mating systems) can influence the variation in animal interactions, but heterogeneity in interactions may arise due to human action, and this has received little attention. We hypothesised that social-mediated differences in the landscape can have consequences for interactions for free-roaming dogs and consequently for disease spread. We used GPS collared data from 31 free-ranging dogs present along a farm-village gradient to build social contact networks to examine differences in interactions. Additionally, we simulated scenarios of rabies infection of two types - dumb and furious - on the contact network structure of village and farm dogs. We found significant differences in the node degree between village and farm contact networks, but the node betweenness was similar. Our simulations showed that infections arising in village dogs are likely to spread to that network, but infections arising in farm dogs get are less likely to spread. Interestingly, this was independent of the type of rabies infection. We discuss our results in light of control strategies for managing rabies in the tropics.