α-Hydroxybutyrate Is an Early Biomarker of Insulin Resistance and Glucose Intolerance in a Nondiabetic Population
Figure 2
Classification of subjects as insulin sensitive or insulin resistant.
Random Forest statistical analysis of biochemical profiling (metabolomic) data. The Importance Plot rank of metabolites according to the contribution of each to the classification of 399 subjects into insulin sensitive (IS, MFFM>45 µmol·min−1·kgFFM−1, top two-thirds of subjects, n = 261) or insulin resistant (IR, MFFM<45 µmol·min−1·kgFFM−1, bottom third of subjects, n = 138) groups. Metabolites are listed on the y-axis in order of importance, decreasing in importance from the top to bottom. The mean decrease in accuracy for each metabolite is plotted on x-axis. INSET: The Confusion Matrix showing the prediction accuracy of the separation of the top two-thirds (IS) from the bottom third (IR) is ∼76%.