Fig. 2: Identification of clusters under the YSC model.
From: Quantitatively defining species boundaries with more efficiency and more biological realism
Top: Validating SNAPPER and StarBeast3 for their abilities to recover clusters from simulated data. Each coloured bar has a 95% Binomial credible interval based on the number of clusters used to estimate its probability. Bottom: Species identified in the gecko (left) and primate (right) datasets, under the YSC model. The clustering scheme presented is the one which occurred with the maximal-posterior probability across all values of threshold ϵ that are labelled with a*. The marginal posterior support is indicated below each cluster (for ϵ = 10−2). Note that the remaining taxa in the bush baby dataset are omitted from the figure because they exist as singleton clusters.
