Data from: Species delimitation with ABC and other coalescent-based methods: a test of accuracy with simulations and an empirical example with lizards of the Liolaemus darwinii complex (Squamata: Liolaemidae)

Species delimitation is a major research focus in evolutionary biology
because accurate species boundaries are a prerequisite for the study of
speciation. New species delimitation methods (SDMs) can accommodate
non-monophyletic species and gene tree discordance as a result of
incomplete lineage sorting via the coalescent model, but do not explicitly
accommodate gene flow after divergence. Approximate Bayesian Computation
(ABC) can incorporate gene flow and estimate other relevant parameters of
the speciation process while testing alternative species delimitation
hypotheses. We evaluated the accuracy of BPP, SpeDeSTEM, and ABC for
delimiting species using simulated data and applied these methods to
empirical data from lizards of the Liolaemus darwinii complex. Overall,
BPP was the most accurate, ABC showed an intermediate accuracy, and
SpeDeSTEM was the least accurate under most simulated conditions. All
three SDMs showed lower accuracy when speciation occurred despite gene
flow, as found in previous studies, but ABC was the method with the
smallest decrease in accuracy. All three SDMs consistently supported the
distinctness of southern and northern lineages within L. darwinii. These
SDMs based on genetic data should be complemented with novel SDMs based on
morphological and ecological data to achieve truly integrative and
statistically robust approaches to species discovery.