Principal Coordinates Analysis
To test for associations between R genes and the phenotypes collected on
the cloned plants across the gradient, we performed simple linear
modeling as well as mixed effect modeling, which treats CMI as a fixed
effect, and source prairie as a random effect. We found one significant
correlation between R Genes of the various populations and leaf blotch.
To interrogate which R genes were responsible for the association
between leaf blotch and R gene counts, we performed a census of which R
gene NBARC sequences occurred in which individuals (Table S5). The
census was then used to generate a principal coordinate analysis in the
R package “vegan” (Oksanen 2015) against each of the five phenotypic
covariates included in the study. P-values were corrected for multiple
tests using the Benjamini & Hochberg method (1995).