Fig. 4

Machine learning findings. A The emerging ten most important bacteria in the machine learning models were compared between those affected with polyps and controls using Wilcoxon test (ns: P > 0.05; *: P ≤ 0.05; **: P ≤ 0.01; ***: P ≤ 0.001; ****: P ≤ 0.0001). B The emerging ten most important species and families according to percentage variable importance, in the stool (left) and tissue (right) machine learning models, respectively. C The correlation matrix showing distinguishing bacterial populations between patients with and without polyposis. D The RF model was used to predict non-polyp mucosa from polyp tissue in affected individuals