Phylogenomic Species Delimitation Dramatically Reduces Species Diversity in an Antarctic Adaptive Radiation

Abstract

Application of genetic data to species delimitation often builds confidence in delimitations previously hypothesized using morphological, ecological, and geographic data and frequently yields recognition of previously-undescribed cryptic diversity. However, a recent critique of genomic data-based species delimitation approaches is that they have the potential to conflate population structure with species diversity, resulting in taxonomic oversplitting. The need for an integrative approach to species delimitation, in which molecular, morphological, ecological, and geographic lines of evidence are evaluated together, is becoming increasingly apparent. Here, we integrate phylogenetic, population genetic, and coalescent analyses of genome-wide sequence data with investigation of variation in multiple morphological traits to delimit species within the Antarctic barbeled plunderfishes (Artedidraconidae: Pogonophryne). Pogonophryne currently comprises 29 valid species, most of which are distinguished solely by variation in ornamentation of the mental barbel that projects from the lower jaw, a structure previously shown to vary widely within a single species. However, our genomic and phenotypic analyses result in a dramatic reduction in the number of distinct species recognized within the clade, providing evidence to support the recognition of no more than six species. We propose to synonymize 24 of the currently recognized species with five species of Pogonophryne. We find genomic and phenotypic evidence for a new species of Pogonophryne from specimens collected in the Ross Sea. Our findings represent a rare example in which application of molecular data provides evidence of taxonomic oversplitting on the basis of morphology, clearly demonstrating the utility of an integrative species delimitation framework.

Genome-wide species delimitation analyses of a silverside fish species complex in central Mexico indicate taxonomic over-splitting

“…Herein, the use of genome-wide data provided an unprecedented resolution that had not been achieved using a scant number of genetic markers to test species delimitation scenarios within the humboldtianum group [11,18]. While species delimitation studies examining thousands of genetic loci often unveil cryptic diversity [22][23][24][25], our study represents one of the few cases where the use of genome-wide SNP data and MSC approaches provide evidence of taxonomic over-splitting [2,5,26]. Recently, it has been recognized that the MSC model can potentially confound population structure with species boundaries-particularly when major sampling gaps near the distribution range exist-leading to an over-estimation of the number of species, a bias compounded by the high statistical power that often results from genome-wide analyses [4][5][6].…”

“…Species delimitation is challenging, as speciation is generally not instantaneous, where there is a diffuse area between populations and species-known as the speciation continuum [1]. The inference of species delimitation thus becomes particularly challenging in recently diverged or closely related species, where it is difficult to differentiate the population-level structure from distinct species as they often undergo incomplete lineage sorting (ILS) or hybridization [2].…”

The mitochondrial genome of the red icefish (Channichthys rugosus) casts doubt on its species status

“…The available molecular data for the genus Channichthys does not allow us to perform a formal genetic species delimitation analysis as was recently done for Pogonophryne (Parker et al, 2022). Nevertheless, a comparison of the divergence between the two Channichthys mitochondrial genomes with the levels of between- and within-species sequence divergence in other notothenioid fishes can inform about the existence of one or several species within the genus.…”

“…Spodareva & Balushkin, 2014). This species richness within Pogonophryne could not be confirmed in a recent genetic analysis, which instead led to the synonymization of 24 out of 29 valid species (Parker, Dornburg, Struthers, Jones, & Near, 2022).…”

Species as a Heuristic: Reconciling Theory and Practice

“…In light of this, the discovery of new specimens, properties, or techniques may lead us to adjust our previous delimitations, but no decision can ever be based on all the potentially relevant or actually available data. The integration of different types of data is therefore key to improving the accuracy of species delimitation, especially in the face of low sampling or high uncertainty ( Pante et al 2015 ; Gratton et al 2016 ; Sukumaran and Knowles 2017 ; Federman et al 2018 ; Hundsdoerfer et al 2019 ; Parker et al 2021 ), consistent with the principles of both the USC and cohesion-based approaches. Decisions about what types of potentially available data or properties should be integrated in these analyses are strategic, in that they will be based on our existing knowledge about the group of organisms in question.…”