Skip to main content
Fig. 4 | BMC Genomics

Fig. 4

From: Machine learning framework for gut microbiome biomarkers discovery and modulation analysis in large-scale obese population

Fig. 4

Identical regional shared biomarkers of obesity. A UpSet plot revealing the amount of filter common species in each discovery cohort and shared by combinations of these datasets. The set size represents the number of biomarkers in each country. The connected dots mean the common biomarkers across connected countries and the number on each column represents the amounts of biomarkers. B The bar plot shows the number of regional shared biomarkers obtained by different methods. The pink color represents the unstable methods. C The Venn plot shows the union of different feature selection methods. D The line plot shows that the results of the Gradient boosted regression trees (GBDT) model have poor stability. The blue line corresponds to the intersection of the result, with the increase in repetition times. E The line plot shows that the AUC of 5 countries with the increase of the regional shared feature numbers by the XGBoost model. F The heatmap shows the ability of different methods to distinguish between obese and healthy individuals in each country and validation cohort

Back to article page