Transmission dynamics of a vector-borne pathogen: An experimental test

Background/Question/Methods

Vector-borne, generalist plant viruses are widespread, yet their transmission dynamics are understudied.  For vector-borne pathogens, transmission scales with the frequency (i.e. proportion) of infected hosts because vectors are actively seeking new hosts. However, at long distances between hosts or in communities that slow vector spread, transmission may instead scale with the density (i.e. count) of infected hosts.   

To determine if host spacing and community context alter the predicted frequency-dependent spread, we studied the transmission of the PAV species of barley yellow dwarf virus (BYDV) among Avena fatua by the aphid Rhopalosiphum padi in a field experiment in central New York State.  We varied the initial frequency (0.11, 0.2, and 0.33) and density (4, 8, and 16) of infected plants, and the spacing between hosts (5, 10, and 20cm) in a fully factorial design.  We replicated this design twice, growing Avena with each of two background hosts of BYDV that vary in their viral and aphid suitability.  We determined the infection status of each Avena plant three times throughout the growing season.  We analyzed transmission over time using logistic regression, with infection status (+/-) as the dependent variable and all main effects and combinations of treatment variables as predictors.

Results/Conclusions

Overall, the data were consistent with frequency-dependent transmission.  BYDV prevalence was highest in the highest initial infection frequency treatment and lowest in the lowest initial infection frequency treatment.  Host community context altered transmission, with the highest levels of BYDV prevalence occurring when Avena was growing with the less suitable background host in the highest initial infection frequency, while all other combinations of initial infection frequency and background host had similar BYDV prevalence.  In addition, we found that some combinations of host spacing and backgrounds differed in their prevalence and incidence of infection.  Finally, we found a significant interaction among initial infection densities, host spacing, and background host species.  For Avena growing with the less suitable background host, BYDV prevalence was more readily predicted by host spacing, while for Avena growing with the more suitable background host, the pattern of BYDV prevalence was more complicated.  Our results are consistent with previous studies of vector-borne pathogens while also showing the importance of considering other epidemiological factors for a more accurate understanding of disease transmission.