Modelling multi-species and multi-mode contact networks: Implications for persistence of bovine tuberculosis at the wildlife–livestock interface

Individual‐ and species‐level heterogeneity in contact rates can alter the ability of a pathogen to invade a host community. Many pathogens have multiple modes of transmission—by direct or indirect contact. It is important to identify the role of heterogeneity in different types of transmission when managing the risk of disease spillover at the interface among different host species. We developed a network‐based analysis to explore how individual‐ and species‐level heterogeneity shape multi‐mode contact networks. We applied this network‐based approach to contact data from proximity loggers collected in a multi‐species host community that contributes to the spillover of the disease bovine tuberculosis (bTB) to cattle populations in Michigan, USA. We used this approach to (a) quantify how individual‐ and species‐level heterogeneity influence direct and indirect contacts in this system, (b) explore how management interventions to control spillovers, such as the installation of deer fences, can alter observed contact networks and c) predict the role that wildlife species have in maintaining bTB in the community. We found that individual‐ and species‐level heterogeneity disproportionately influenced indirect and direct contact networks, with individual‐level heterogeneity having a greater effect on indirect contact networks and species‐level heterogeneity having a greater effect on direct contact networks. We also found that the installation of deer fences significantly reduced deer‐specific indirect contacts. We used the results from our network analysis to show that white‐tailed deer (Odocoileus virginianus) could act as the sole reservoir host for bTB in this community with important implications for understanding past bTB dynamics and managing the persistence of bTB in the future. Synthesis and applications. Analyses of epidemiological networks rarely account for multiple modes of contact, which can lead to an incomplete understanding of how individual‐ and species‐level heterogeneity affect disease transmission. The multi‐mode, multi‐species network analysis we develop in this study illustrates that individual‐ and species‐level heterogeneity can play significantly different roles depending on the type of contact network considered. This has important implications when managing disease at the wildlife–livestock interface, where strategies may need to be multi‐pronged to account for the variable role of heterogeneity on different modes of contact. Analyses