Human disturbances of intact tropical habitats combine with deep coevolution of diverse faunas and their pathogens to create the perfect storm for infectious diseases that are emerging and capturing headlines today. Pathogens possess unique vectors and features of infection, hence it is not easy to determine whether humans are perturbing a tightly evolved system, or whether humans function as just another victim of infection in an ecological dynamic. Indeed, most efforts to identify the source of zoonoses occur after humans are infected and it is difficult to predict zoonoses because spillover events are rare and infection prevalence in reservoir species can be low. We attempt to develop a predictive approach by studying Bartonella bacteria, which are prevalent in many mammalian taxa and are responsible for numerous human infections. Using a phylogenetic approach, we show that humans have frequently been exposed to Bartonella, most recently through spillover from commensal rodents in the genus Rattus. In general, rodents and domestic animals serve as the reservoirs or at least key proximate hosts for most Bartonella genotypes in humans and also in wild animals. Thus despite a long coevolution of Bartonella strains and their particular mammalian hosts, the recent expansion of humans and their commensals has left an imprint on these relationships so that neither geographic origin nor phylogenetic distance help to identify which pathogens may spillover into humans. Instead, a better indicator of future zoonoses may be found by investigating our own backyard.