Joint interactions with humans pose a greater risk of zooanthroponotic disease outbreaks than social interactions among (peri)urban wildlife populations

In this study, we combined social network analysis with epidemiological Susceptible Infected Recovered (SIR) models, to investigate the effects of macaques' (<i>Macaca</i> spp) co-occurrence and joint interactions with humans, and their social grooming of conspecifics, on simulated disease outbreaks. On each of 10 macaque groups representing three different species living in anthropogenic (peri)urban environments in India and Malaysia, we collected behavioral data on macaques' tendencies to co-occur and jointly engage in interactions (e.g. aggression, provisioning, snatching food) with humans within the same time and space, and their social grooming (cleaning and manipulating the fur of another individual) of their conspecifics (see our previous publications: Balasubramaniam et al., 2020, 2021). The data were used to construct both 'human co-interaction networks' and 'social grooming networks'. For each group and network type, we then ran 5000 SIR model simulations. These were run for pathogens of varying transmissibility, by randomly selecting a macaque to be infected first. The model simulation then proceeds until the pathogen reaches saturation or extinction. It does so by allowing animals to move through animals moving across disease states based on dynamic probabilities that are functions of both their network connectedness and pathogen transmissibility.<br>The dataset uploaded is the outcome of these simulations. For each "first infected macaque", we have included the average outbreak size from across the simulations in which it was the first infected individual. This information has been paired with various sociodemographic (e.g. group size, macaque species, sex of the individual) and behavioral characteristics (e.g. their direct and indirect connectedness or centrality within their human co-interaction networks and grooming networks, dominance ranks, and anthropogenic interactions indicated by their frequencies of interactions with humans and proportions of time spent foraging on human foods). The dataset was used to run GLMMs aimed at examining the effects of the above sociodemographic attributes and behavioral traits of first-infected individuals on the outcome of model-predicted disease outbreak sizes. Thus, we have also provided a second file that contains R code for the GLMMs.<br>The datasets provided were used to write a manuscript that is currently under review in the journal <i>Scientific Reports. </i>They have not been published elsewhere, or made available elsewhere.