Fig. 3: RAC and Silhouette score plots for N-ICE dataset.
From: Definition of the microbial rare biosphere through unsupervised machine learning
a Relative abundance of each phylogenetic unit (either ASV, OTU or mOTU) from all samples. For each sample, phylogenetic units were ordered in the x axis from the most to the least abundant. The y axis shows the abundance score of each phylogenetic unit in Log10 scale. For context, three relative abundance thresholds are highlighted with dashed lines (0.01%, 0.10% and 1.00%). b Silhouette scores obtained for each phylogenetic unit from all samples. Phylogenetic units are ordered from highest to lowest Silhouette score. For (a, b), lines were used to group phylogenetic units in the same sample and abundance classification. Default classifications (“rare”, “undetermined” and “abundant”) were colored coded.
