Fig. 4From: Performance determinants of unsupervised clustering methods for microbiome dataIllustration of branching process, abundance sum of OTUs with high abundance plot, Rand indexes plot, and PCoA plot of MartÃnez dataset with descendants. A A schematic of the branching process, where sequences of each OTU were randomly assigned to either of the two daughter branches. B Abundance sum of OTUs with high mean abundance plot of the MartÃnez Dataset with first, second, and third generation descendants. Lines in the bar plot indicate the 95th percentile intervals of the Rand indices from 200 repeated simulations. C Rand indices of the MartÃnez dataset with the original dataset and simulations of the first, second, and third generation descendants. Lines in the bar plot indicate the 95th percentile intervals of the Rand indices from 200 repeated simulations. D Bray-Curtis beta diversity PCoA of the original MartÃnez dataset, as well as examples of datasets with the addition of the first, second, and third generation novel OTUsBack to article page