Skip to main content
Fig. 4 | Environmental Microbiome

Fig. 4

From: Testing the stress gradient hypothesis in soil bacterial communities associated with vegetation belts in the Andean Atacama Desert

Fig. 4

Best discriminant factors to differentiate between the BS and the RSS samples using machine learning. a 10CV accuracy of random forest classification models to discriminate between BS and RSS samples based on the relative abundance of OTUs or chemical component concentrations in different soil compartments at different taxonomic levels. b OTUs ranked from top to bottom, according to their importance in the random forest model (in terms of mean decrease in accuracy (MDA))- c chemical components

Back to article page