A vector–host model for coinfection by barley/cereal yellow dwarf virus

Background/Question/Methods

Interactions between pathogen species within a host can have important effects on disease dynamics, triggering cross protective immunity, synergistic mortality, or altering host resistance, infectivity, and transmission. Here we investigate a mathematical model for a well-studied multi-pathogen, multi-vector system, barley and cereal yellow dwarf viruses (B/CYDV). B/CYDV is a suite of aphid-vectored pathogens that affect diverse host communities, including economically-important crops such as barley, wheat, and oats. Coinfection by multiple strains of B/CYDV can result in elevated virulence, incidence, and transmission rates.

We develop a model for a single host, two pathogen strains and n vector species, consisting of a system of nonlinear ordinary differential equations. A single parameter describes the degree of relatedness of the strains and the amount of cross-protection between them. This model allows us to compute the threshold quantities governing the endemicity of infection and coinfection and determine the relative importance of each parameter on the system dynamics. In addition, we consider both exponential and logistic vector growth functions to determine the effect of density dependence in the vector.

Results/Conclusions

We compute the basic and type reproduction numbers of the model and demonstrate that although the basic reproduction number describes stability of the disease-free equilibrium, the type reproduction numbers better describe the individual dynamics of each strain and of coinfection. We then conduct a numerical sensitivity analysis on the components of the endemic equilibrium. Disease transmission rates and vector birth and mortality rates are the most influential parameters on the equilibrium prevalences of infection and coinfection. The vector growth parameters become extremely sensitive under the assumption of a logistic vector growth curve.

Our results suggest that controlling or eradicating B/CYDV may be best accomplished by reducing the vector population and its rate of movement between hosts. In addition, empirical determination of the degree and form of vector density dependence will be critical for effective predictions about coinfection in natural host populations.