The roles of multi-level selection and recombination in maintaining genetic diversity in an insect virus

Background/Question/Methods

Though selection in a host-pathogen system typically acts on two distinct scales (i.e. within-host and between-host), models that incorporate these scales are difficult to apply to real systems, because within-host pathogen dynamics are poorly understood.  Furthermore, relating selection on phenotypes to changes in genotype frequencies is often difficult.  With the gypsy moth caterpillar and its associated baculovirus, however, we can overcome these obstacles and thus model pathogen evolution in a real system.  We began by fitting a suite of biologically-motivated birth-death models to dose-response data.  Using AIC, we then selected the model that best described the within-host dynamics.  With this model, we fit the parameters of 96 virus strains to another dose-response dataset.  Having recently received whole-genome sequence data, we will soon assign effect sizes on parameter estimates to each polymorphic locus.  Concurrently, we built a multi-generational model that incorporates within-host and between-host dynamics.  After incorporating our sequence data, we will use this model to see the effects of recombination, selection and drift on the evolution of the pathogen population.

Results/Conclusions

Preliminary analysis of sequence data revealed three features of the pathogen population: high levels of genetic polymorphisms, high levels of co-infections, and low levels of linkage disequilibrium.  These features suggest that recombination plays an important role in the evolution of this pathogen.  Since the parameter estimation of our 96 virus strains revealed high levels of phenotypic polymorphism, our next step is to map the polymorphisms to genetic loci.  With these data we will then incorporate realistic recombination into our multi-generational simulation.  Meanwhile, preliminary results from our multi-generational simulations show that a tradeoff exists between within-host and between-host transmission.  We thus expect different virus genotypes to be favored in different situations (e.g. co-infection vs. solo infection, high host density vs. low host density).  Furthermore, existing theory and preliminary output from our model suggest that these tradeoffs can potentially lead to high levels of genetic polymorphism.  Ultimately, we expect to show that multi-level selection, in the presence of high recombination rates, can explain the high levels of genetic diversity found in this pathogen system.