A matrix model for density-dependent selection in stage-classified populations, with application to pesticide resistance in Tribolium

The study of eco-evolutionary dynamics is based on the idea that ecological and evolutionary processes may operate on the same, or very similar, time scales, and that interactions of ecological and evolutionary processes may have important consequences. Here we develop a model that combines Mendelian population genetics with nonlinear demography to create a truly eco-evolutionary model. We use the vec-permutation matrix approach, classifying individuals by stage and genotype. The demographic component is female dominant and density-dependent. The genetic component includes random mating by stage and genotype, and arbitrary effects of genotype on the demographic phenotype. Mutation is neglected. The result is a nonlinear matrix population model that projects the stage × genotype distribution. We show that the results can include bifurcations of population dynamics driven by the response to selection. We present analytical criteria that determine whether one allele excludes the other or if they persist in a protected polymorphism. The results are based on local stability analysis of the homozygous boundary equilibria.
As an example, we use a density-dependent stage-classified model of the flour beetle Tribolium castaneum. Our model permits arbitrary life-cycle complexity and nonlinearity. Tribolium has developed resistance to the pesticide malathion due to a dominant allele at a single autosomal locus. Using parameters reported from laboratory experiments, we show that the model successfully describes the dynamics of both resistant and susceptible homozygotes, and the outcome of a selection experiment containing both alleles. Stability analysis of the boundary equilibria confirms that the resistant allele excludes the susceptible allele, even in the absence of malathion, agreeing with previously reported results.