Skip to main content
Fig. 5 | BMC Genomics

Fig. 5

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

Fig. 5

The plot shows the biomarkers identified by counterfactual inference. A The heatmap graph at the bottom shows the median abundance of regional shared biomarkers in the Chinese healthy individuals (CHN_H), a random Chinese original obese sample (Orig_OB), and the corresponding counterfactual instance (CF), respectively. The red-marked biomarkers mean that the abundance is increased in the counterfactual instance; the blue-marked biomarkers mean that the abundance is reduced in the counterfactual instance. The heatmap graph at the top shows the general direction of biomarkers across different countries by pooling all the counterfactual instances. The red color means that the median abundance of the biomarkers is raised in the counterfactual instances; in contrast, the blue color means that the median abundance of the biomarkers is decreased in the counterfactual instances. The gray color means that the difference in median abundance of the biomarkers between the original obese individual and the counterfactual instance is less than 0.1. B-C The box plots (box limits, upper and lower quartiles; center line, median; whiskers, 1.5 × interquartile range) show the abundance of biomarkers in obese and healthy people across different countries

Back to article page