Impact of Model Violations on the Inference of Species Boundaries Under the Multispecies Coalescent

Abstract

The use of genetic data for identifying species-level lineages across the tree of life has received increasing attention in the field of systematics over the past decade. The multispecies coalescent model provides a framework for understanding the process of lineage divergence and has become widely adopted for delimiting species. However, because these studies lack an explicit assessment of model fit, in many cases, the accuracy of the inferred species boundaries are unknown. This is concerning given the large amount of empirical data and theory that highlight the complexity of the speciation process. Here, we seek to fill this gap by using simulation to characterize the sensitivity of inference under the multispecies coalescent (MSC) to several violations of model assumptions thought to be common in empirical data. We also assess the fit of the MSC model to empirical data in the context of species delimitation. Our results show substantial variation in model fit across data sets. Posterior predictive tests find the poorest model performance in data sets that were hypothesized to be impacted by model violations. We also show that while the inferences assuming the MSC are robust to minor model violations, such inferences can be biased under some biologically plausible scenarios. Taken together, these results suggest that researchers can identify individual data sets in which species delimitation under the MSC is likely to be problematic, thereby highlighting the cases where additional lines of evidence to identify species boundaries are particularly important to collect. Our study supports a growing body of work highlighting the importance of model checking in phylogenetics, and the usefulness of tailoring tests of model fit to assess the reliability of particular inferences. [Populations structure, gene flow, demographic changes, posterior prediction, simulation, genetics.].

Defining Relictual Biodiversity: Conservation Units in Speckled Dace (Cyprinidae:Rhinichthys osculus) of the Greater Death Valley Ecosystem

“…Two approaches were employed to determine the number of discrete conservation units in the 278 Death Valley region: Multispecies coalescent (MSC) methods and unsupervised machine learning (UML) algorithms. Machine learning algorithms have been proposed as alternatives to 280 MSC methods, which seemingly over-split taxa under certain conditions (Barley, Brown, & Thomson, 2018;Leaché, Zhu, Rannala, & Yang, 2018), and parse populations rather than 282 speciation events (Sukumaran & Knowles, 2017). This stems from the many assumptions implicit to MSC, such as random mating, neutral markers, a lack of post-speciation gene flow, 284 and no within-locus recombination or linkage disequilibrium (Degnan & Rosenberg, 2009).…”

Model-based species delimitation: are coalescent species reproductively isolated?

“…However, further work is needed to move beyond such speculation. Overall, our 605 results highlight the value of considering the processes that underlie speciation, in conjunction 606 with the assumptions of each species delimitation method employed, as a fruitful means for 607 understanding conflicts (see also Barley et al 2018;Smith and Carstens 2018). While reproductive isolation has long been a cornerstone of speciation research, it can rarely be 617 used as an operational delimitation method because collecting such data is not tractable for most 618 systems.…”

Species Delimitation in Herpetology

“…In addition, many MSC-based methods also make unrealistic assumptions about the population structure, constant rates of evolution, and methods for reconstructing individual genes. Assumptions about these and other factors rarely are tested or considered in applications of MSC studies (Zhang et al, 2011;Olave et al, 2014;Barley et al, 2018). Limited sampling of specimens is especially likely to result in clusters of geographically proximate samples being identified as ''species'' (Schwartz and McKelvey, 2008;Rittmeyer and Austin, 2012;Barley et al, 2018), even if the only genetic structure within the samples represents classic isolation by distance (Wright, 1943;Slatkin and Maddison, 1990).…”

“…In some cases, wide-ranging species have been split into clusters of geographically proximate samples that are re-named as species, even though genetic analyses of adjacent populations suggest continuous gene flow across the range of samples. MSC-based methods often support a split of a continuous geographic cline into multiple species (Barley et al, 2018). For example, Burbrink and Guiher (2015) proposed splitting both Agkistrodon contortrix and Agkistrodon piscivorous into two species each, dividing eastern and western populations of both species in the middle of the species range, with broad geographic regions of ''hybrids'' on either side of the putative species boundaries.…”