Fig. 5.

Clinical outcomes classification based on the signatures of gut microbiome and metabolome. Random forest classifiers composed of bacteria, metabolites and their combination were constructed to discriminate patients with DCB and NDB groups. The selected features (A), receiver operating characteristic (ROC) curve in train and test sets (B), and the MeanDecreaseAccuracy of selected features (C) based on bacteria (species level). The selected features (D), ROC curve in train and test sets (E), and the MeanDecreaseAccuracy of selected features (F) based on metabolites. The selected features (G), ROC curve in train and test sets (H), and the MeanDecreaseAccuracy of selected features (I) based on combination bacteria and metabolites