Fig. 4From: Fecal source identification using random forestDistribution of amplicon sequence variants (ASVs) selected among the Clostridiales classifiers. (a) Mean and distribution of the number of ASVs belonging to the different classifiers for each source of fecal samples. (b) Heatmap representing the relative abundance of the ASVs selected within the eight classifiers (represented on the horizontal axis) within the samples (listed on the right) used to build the classifiers. Samples were clustered using the UPGMA algorithm based on the Bray–Curtis dissimilarity matrixBack to article page