Landscape structure as a driver of eco-evolution in host-parasite systems.

Spatial network structure of biological systems drives ecology and evolution by distributing organisms and their genes. The ubiquitous host-parasite systems are no exception. However, past theoretical work has largely focused on simple spatial structures, such as grids, hampering the translation of theoretical predictions to real ecosystems. Thus, we develop an eco-evolutionary metapopulation model of host-parasite dynamics where hosts and parasites disperse through realistically complex spatial networks representing major biomes using river-like and terrestrial-like networks. We generate the testable prediction that parasite virulence, or how parasites harm their hosts, peaks at intermediate dispersal values in river-like systems while it increases with increasing dispersal in terrestrial-like systems. In river-like systems, virulence also reaches higher overall values. Moreover, we show that kin selection is the main driver of virulence evolution. Spatial networks generate characteristic patterns of parasite relatedness which drive differential virulence evolution. Finally, we show that accounting for virulence evolution allows us to predict the distribution of key epidemiological variables (e.g., parasite extinction risks) within spatial networks. Our study highlights how eco-evolutionary feedbacks can be understood in light of spatial network structure by linking network topology to classical evolutionary mechanisms such as kin selection.