Speciation due to hybrid necrosis in plant-pathogen models
We develop a model for speciation due to postzygotic incompatibility generated by autoimmune reactions. The model is based on predator-prey interactions between a host plants and their pathogens. Such interactions are often frequency-dependent, so that pathogen attack is focused on the most abundant plant phenotype, while rare plant types may escape pathogen attack. Thus, frequency dependence can generate disruptive selection, which can give rise to speciation if distant phenotypes become reproductively isolated. Based on recent experimental evidence from {\it Arabidopsis}, we assume that at the molecular level, incompatibility between strains is caused by epistatic interactions between two proteins in the plant immune system, the guard and the guardee. Within each plant strain, immune reactions occur when the guardee protein is modified by a pathogen effector, and the guard subsequently binds to the guardee, thus precipitating an immune response. However, when guard and guardee proteins come from phenotypically distant parents, a hybrid's immune system can be triggered by erroneous interactions between these proteins even in the absence of pathogen attack, leading to severe autoimmune reactions in hybrids. Our model shows how phenotypic variation generated by frequency-dependent host-pathogen interactions can lead to postzygotic incompatibility between extremal types, and hence to speciation.