Inferring host range dynamics from comparative data: the protozoan parasites of new world monkeys

Uncovering the ecological determinants of parasite host range is a central goal of comparative parasitology and infectious disease ecology. But while parasites are often distributed nonrandomly across the host phylogeny, such patterns are difficult to interpret without a genealogy for the parasite samples and without knowing what sorts of ecological dynamics might lead to what sorts of nonrandomness. We investigated inferences from comparative data, using presence/absence records from protozoan parasites of the New World monkeys. We first demonstrate several distinct types of phylogenetic signal in these data, showing, for example, that parasite species are clustered on the host tree and that closely related host species harbor similar numbers of parasite species. We then show that all of these patterns can be generated by a single, simple dynamical model, in which parasite host range changes more rapidly than host speciation/extinction and parasites preferentially colonize uninfected host species that are closely related to their existing hosts. Fitting this model to data, we then estimate its parameters. Finally, we caution that quite different ecological processes can lead to similar signatures but show how phylogenetic variation in host susceptibility can be distinguished from a tendency for parasites to colonize closely related hosts. Our new process-based analyses, which estimate meaningful parameters, should be useful for inferring the determinants of parasite host range and transmission success.