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

Abstract

22The tips in the tree of life serve as foci for conservation and management, yet clear delimitations are masked by inherent variance at the species-population interface. Analyses using thousands of 24 nuclear loci can potentially sort inconsistencies, yet standard categories applied to this parsing are themselves potentially conflicting and/or subjective [e.g., DPS (distinct population 26 segments); DUs (Diagnosable Units-Canada); MUs (management units); SSP (subspecies); Evolutionarily Significant Units (ESUs)]. One potential solution for consistent categorization is 28 to create a comparative framework by accumulating statistical results from independent studies and evaluating congruence among data sets. Our study illustrates this approach in speckled dace 30 (Cyprinidae: Rhinichthys osculus) endemic to two basins (Owens and Amargosa) in the Death Valley ecosystem (DVE). These fish persist in the Mojave Desert as isolated Pleistocene-relicts 32 and are of conservation concern, but lack formal taxonomic descriptions/designations. Doubledigest RAD (ddRAD) methods identified 14,355 SNP loci across 10 populations (N=140). 34Species delimitation analyses [multispecies coalescent (MSC) and unsupervised machine learning (UML)] delineated four putative ESUs. F ST outlier loci (N=106) were juxtaposed to 36 uncover the potential for localized adaptations. We detected one hybrid population that resulted from upstream reconnection of habitat following contemporary pluvial periods, whereas 38 remaining populations represent relics of ancient tectonism within geographically-isolated springs and groundwater-fed streams. Our study offers three salient conclusions: A blueprint for 40 a multi-faceted delimitation of conservation units; a proposed mechanism by which criteria for intraspecific biodiversity can be potentially standardized; and a strong argument for the proactive 42 management of critically-endangered DVE fishes.3

The choices we make and the impacts they have: Machine learning and species delimitation in North American box turtles (Terrapene spp.)

“…By comparison, BFD* partitioned all groups, which may reflect a vulnerability to local structure at the population level, as reported by others for MSC methods (Sukumaran & Knowles, 2017). BFD* and VAE partitioned equally in Mussmann et al (2020), although their populations were relictual and without contemporary connectivity, whereas Terrapene reflects both historical (Figure 3d) and contemporary gene flow (Martin et al, 2020). In corroboration, other studies have also demonstrated reticulation to condense VAE clusters (Derkarabetian et al, 2019;Newton et al, 2020).…”

“…As with prior studies (Derkarabetian et al, 2019;Mussmann et al, 2020), we also found considerable variation among methods, some of which can be attributed either to idiosyncrasies in the data or to algorithms and their implementation. First, among RF methods…”

“…In contrast to Derkarabetian et al (2019), we found t-SNE the most inclined to produce inconsistent groupings, a pattern most prevalent with the gap statistic (Supporting Information Appendix B1-B60). Mussmann et al (2020) concurred, although in their case it was PAM + HMSW. We see this as an inherent problem relating to data structure.…”

“…Although phylogenetic structuring was present in some analyses (e.g., Figure 2), it is insufficient to mandate recognition beyond the subspecific level. However, special guidelines that delineate relictual lineages may be warranted (Mussmann et al, 2020), particularly…”

“…Potentials include: Data quantity (Newton et al, 2020), the proportion of missing data (Mussmann et al, 2020), and evolutionary complexity (Austerlitz et al, 2009). Here, we employ M-L algorithms alongside coalescent methods such as BFD* (Leaché et al, 2014) as vehicles to parse a taxonomically recalcitrant clade.…”