Gut microbiota predicts the diagnosis of ulcerative colitis in Saudi children

Figure 1 The predictive power of fecal bacteriome. A: Area under the curve (AUC) based on the entire bacterial species shows 89.5% (79.1%-100%CI) accuracy in predicting ulcerative colitis (UC); B: Random forest algorithm was performed on the entire dataset to identify important features significantly predictive of UC increased the AUC to 97.6% (94.2-100%CI).

Figure 2 The predictive power of fecal bacteriophage. A: Area under the curve (AUC) based on the entire bacteriophage shows of 87.4% (75.9%-98.8%) in predicting ulcerative colitis (UC) in stool samples; B: Random forest algorithm was performed on the entire dataset to identify important features significantly predictive of UC increased the AUC to 94.5% (87.8%-100%CI).