Figure 5

Multivariate exploratory ROC curve analysis showing 25 potential metabolite biomarkers selected by applying support vector machine (SVM), partial least squares discriminant analysis (PLS-DA), and random forests (a); AUC curve with 1 value showing high predictability of the model at 100 features (model 6) in MetaboAnalyst5.0 (b); and annotated up regulated metabolite biomarkers apigenin-7-glucoside (c), uridine (d), adenosyl homocysteine (e), cyclic guanosine monophosphate (cGMP) (g), tyrosine (h), pantothenic acid (i), riboflavin (l) and down regulated metabolites homocysteine (f), ajmaline (j) and adenosine (k).