Our preliminary experimental data reveals that coinfection has exceptionally strong impacts on pathogen performance in Arabidopsis. To test for these effects, we conducted infections consisting of a luciferase-labeled focal strain of P. viridiflava that was separately co-inoculated with each of 60 randomly chosen strains from the P. syringae complex. At 36 hours post-inoculation, luciferase activity (i.e., photon counts) in each infected plant was measured to quantify abundance of the focal strain. This was replicated in triplicate for two different focal strains. Four aspects of these results are favorable for the proposed work:
- First, the effects of coinfection on pathogen performance are large. The mean abundance of both focal strains differed by two orders of magnitude between the most and least favorable coinfection combinations.
- Second, these effects are highly consistent and dwarf experimental noise. The identity of the coinfecting strain explained over 70% of the variance in focal strain abundance in our experiments (linear mixed model, abundance ~ coinfecting strain + batch effect covariates), and this effect was statistically significant (P < 1e-16).
- Third, our data indicate the potential for both costs and benefits to coinfection, depending on the identity of the coinfecting strains. In 40% of the pairwise coinfection combinations, the focal strain grew to a higher abundance than when singly inoculated; conversely, in 60% of cases, its abundance decreased relative to single infections.
- Fourth, the two focal strains differed in how their abundance was affected by the coinfecting strains (aforementioned linear mixed model; focal strain x coinfecting strain interaction, P < 0.001). This underscores the importance of accounting for genotype x genotype interactions, as we propose to do, when predicting infection outcomes.