Fig. 5

Global biogeographic patterns of ATGs. a, b Global distribution maps of ATG cluster diversity (a) and abundance (b). Global ATG cluster diversity and abundance were predicted based on a random forest model using 97 spatial covariates. Four-fifths of the samples were used as the training set, while one-fifth served as the testing set (for cluster diversity, training set tenfold cross-validation R² = 0.691, testing set R² = 0.587; for abundance, training set tenfold cross-validation R² = 0.577, testing set R² = 0.501; Supplementary Fig. S9). c, d Latitudinal distribution of global ATG cluster diversity (c) and abundance (d). The dashed lines represent the average of global ATG cluster diversity and abundance. e, f Relative importance of each category of variables for predicting ATG cluster diversity (e) and abundance (f). ANT, anthropogenic; TEM, temperature; SP, soil properties; LUAO, land use and others; OCV, other climatic variables; RAD, radiation; PRE, precipitation; MOI, moisture