Pearson correlation coefficients obtained for each variable pair (n = 11 009 species).

The blue color is related with positive values, while red color represents negative correlation (not evinced in this study). Encircled represents all the coefficients that are superior to the threshold declared in materials and methods (0.9). By trimming one of the variables of every encircled pair, the data set was prepared for constructing xgboost model 2.