predicted_trait_data.csv

Trait data used in our analyses. Data included trait values that could be directly associated to taxa in our study based on matching Latin binomials, and trait value estimates based on hidden state prediction using weighted square-change parsimony. Because we converted all trait values to numeric (e.g., Gram-negative = 0 and Gram-positive = 1), state predictions for initially discrete traits were allowed to be fractional (e.g., a Gram-positive score of 0.5), reflecting their probabilistic uncertainty.