Figure 8.
Random forest regression between microbiota and genetic distance to tomato suggests that phylogenetically related plants harbor similar soil microbial communities. Fivefold cross‐validation was used to train/predict log10 genetic distance to tomato as a function of soil microbiota composition across all plant species samples collected in this study. The scatterplot shows the relationship between true and predicted genetic distance. The gray line indicates the linear regression trend line between true and predicted values (Pearson R 2 = .52, p < .001; mean squared error = 0.21). The dotted line marks the ideal 1:1 relationship between true and predicted values