Predicting the global mammalian viral sharing network using phylogeography

Understanding interspecific viral transmission is key to understanding viral ecology and evolution, disease spillover into humans, and the consequences of global change. Prior work has demonstrated that macroecological factors drive viral sharing in some mammalian groups, but analyses have never attempted to predict viral sharing in a pan-mammalian context. Here we show that host phylogenetic similarity and geographic range overlap are strong, nonlinear predictors of viral sharing among species across the entire mammal class. Using these traits, we predict global viral sharing patterns across 4196 mammal species and show that our simulated network successfully predicts viral sharing and reservoir host status using internal validation and an external dataset. We predict high rates of mammalian viral sharing in the tropics, particularly among rodents and bats, and that within- and between-order sharing differs geographically and taxonomically. Our results emphasize the importance of macroecological factors in shaping mammalian viral communities, and provide a robust, general model to predict viral host range and guide pathogen surveillance and conservation efforts.