Biology subjects

Pyron, R. A.

Publications and source records attributed to Pyron, R. A..

7 recordsLinked to original sources

bioRxiv2026

Fully integrative species delimitation with machine learning and diverse data types in delimSOM 2.0

Species delimitation increasingly relies on multiple sources of evidence, but most frameworks are still limited to genetics and morphology or analyze data types separately and compare results qualitatively. We present delimSOM 2.0, an R package that leverages unsupervised machine learning through multilayer self-organizing maps. Any kind of input data, including genomic, phenotypic, ecological, and spatial, is analyzed jointly, retaining each data type as a separate layer and balancing their contributions through layer-specific weighting. Our package provides functions for preprocessing, replicate training, multivariate clustering, hyperparameter optimization, model-quality assessment, and variable- and layer-importance analyses. We evaluated our approach using two simulation sets and eight empirical studies spanning diverse taxa and data types. Overall, performance was strongest with larger sample sizes (>10 individuals per lineage) and modest numbers of lineages (K < 5), but unlike most species delimitation methods, it can reliably infer K = 1. Across two-population demographic simulations of genetic data, our method performed similarly to the well-established DAPC across gene-flow and sampling regimes, was more conservative than STRUCTURE and sNMF in inferring K = 2 and showed mixed support for K = 1 and K = 2 at intermediate differentiation levels and with gene flow. Crucially, our method is robust to missing data and variation in hyperparameters. Empirical examples showed that genetic data were typically most influential, but phenotypic, ecological, and spatial layers often contributed substantial complementary information. By combining diverse data types into a single unsupervised analysis, delimSOM 2.0 enables fully integrative species delimitation, reduces over-splitting, and advances efforts toward more objective and standardized taxonomy and speciation biology.

evolutionary biology↗

bioRxiv2025

Habitat and complex life cycles promote morphological diversity in salamander limb bones

Salamander evolution featured multiple transitions between water and land that likely promoted distinct adaptations in limb bones for buoyancy control versus increased load-bearing capacities, respectively. Many extant species spend their entire lives either in water or on land, while, others undergo water-land transitions within their lifetime. However, exposure to both environments may impose competing demands that restrict adaptive evolution for a particular habitat. Using a 3D morphological dataset of 133 species spanning the phylogenetic and ecological breadth of salamanders, we find that the external and internal morphology of limb bones are evolutionarily decoupled, which increases the evolvability of limb bones in response to diverse mechanical demands. Terrestrial salamanders have stiffer bones with greater resistance to fracture, while aquatic species have denser bones that are hypothesized to aid in buoyancy regulation. We uncover a functional trade-off between stiffness and density that promotes stiff yet lightweight bones in terrestrial lineages. Released from terrestrial constraints, aquatic paedomorphs have disparate external morphologies, whereas terrestrial direct developers consistently share a rod-like bone shape. Aquatic and terrestrial multiphasic taxa show less morphological divergence than monophasic species living in comparable habitats but are not constrained by their complex life cycle. Multiphasic species have distinct external limb bone shapes that have evolved as fast or faster than monophasic species. Taken together, we propose that the trade-offs imposed by different habitats and complex life cycles increase limb bone diversity by promoting alternate evolutionary pathways.

evolutionary biology↗

bioRxiv2025

Telomere-to-telomere reference genome of the common five-lined skink, Plestiodon fasciatus (Squamata: Scincidae)

Although the publication of high-quality reference genomes is steadily increasing, many clades remain chronically neglected. Skinks (order: Squamata; family: Scincidae) are one of the most diverse lizard families (1,785 species), yet there are currently just six published chromosome-level skink genomes. Here, we present the first telomere-to-telomere, chromosome-level reference genome for one of the most abundant lizards in the eastern United States, the common five-lined skink (Plestiodon fasciatus). Through the sequencing of RNA, long-read DNA, and Hi-C chromatin interactions, we produced an annotated reference genome (N50 = 227MB, L50 = 3) consisting of 6 macrochromosome pairs and 7 microchromosome pairs with 98% of BUSCO genes represented (lineage: Sauropsida; 7480 BUSCO markers) represented, providing one of the most complete skink genomes to date: rPleFas1.1. Functional annotation predicts 32,520 protein-coding genes (16,100 unique, named genes) with an average gene length of 9,372bp. Repeat annotations estimate that transposable elements comprise 46.7% of the genome, for which we show the amount and content is remarkably conserved across Scincidae.

genomics↗

bioRxiv2024

The first complete assembly for a lungless urodelan with a "miniaturized" genome, the Northern Dusky Salamander (Plethodontidae: Desmognathus fuscus)

Salamanders have some of the largest genomes among animals. However, there is a great disparity in total size, ranging from [~]8-120GB depending on the lineage. Species in the lungless genus Desmognathus (Plethodontidae) are among the smallest, with estimated genome sizes of 13-15GB. Salamander genomes have exceptional interest in numerous topics ranging from genome-size evolution, the genetic basis of evolutionary differences in life history, and the physiological basis of regeneration, vision, and immunity. However, their large genomes have limited previous attempts at sequencing and assembly, particularly given the difficulties of mapping extensive repeat regions with short-read data. Here we assemble a draft genome of Desmognathus fuscus using PacBio HiFi reads and generate transcriptomic data from two specimens. The combined assembly resulted in a 16.1GB genome in 19,640 contigs and an N/L50 of 2,455/1.75MB, with the longest contig at 27.9MB. The assembly and transcriptome are nearly complete with 93% of the 5,310 BUSCO tetrapod orthologs identified. Attempts to scaffold these data to the existing Ambystoma genome resulted in only 5.8GB of the D. fuscus genome mapping to this reference. This low success suggests substantial syntenic and sequence divergence across salamanders, which may be the result of significant miniaturization in the Desmognathus genome. Identification and annotation of transposable elements reveals that only 26% of the genome is single copy with 74% corresponding to TEs. The most common class of TE in the genome are LTRs (35% of the total genome) and LINEs ([~]15% of the genome). The relative divergence landscape of these TEs shows an early expansion and slow contraction of LINEs, followed by a quick recent expansion of both LTRs and DNA transposons. This assembly will serve as an important reference for amphibian genomics.

genomics↗

bioRxiv2024

Complex Hybridization in a Clade of Polytypic Salamanders (Plethodontidae: Desmognathus) Uncovered by Estimating Higher-Level Phylogenetic Networks

Hybridization between incipient lineages is a common feature of ecomorphological diversification. We examine these phenomena in the Pisgah clade of Desmognathus salamanders from the southern Appalachian Mountains of the eastern United States. The group contains four to seven species exhibiting two discrete phenotypes, aquatic "shovel-nosed" and semi-aquatic "black-bellied" forms. These ecomorphologies are ancient and have apparently been transmitted repeatedly between lineages through introgression. Geographically proximate populations of both phenotypes exhibit admixture, and at least two black-bellied lineages have been produced via reticulations between shovel-nosed parentals, suggesting complex transmission dynamics. However, computational constraints currently limit our ability to reconstruct network radiations from gene-tree data. Available methods are limited to level-1 networks wherein reticulations do not share edges, and higher-level networks may be non-identifiable in many cases. We present a heuristic approach to recover information from higher-level networks across a range of potentially identifiable empirical scenarios, supported by theory and simulation. When extrinsic information indicating the location and direction of hybridization events is available, our method can yield successful estimates of non-level-1 networks, or at least a reduced possible set thereof. Phylogenomic data strongly support a single backbone topology with up to five overlapping hybrid edges. These results suggest an unusual mechanism of ecomorphological hybrid speciation, wherein a binary threshold trait causes hybrids to shift between two microhabitat niches, promoting ecological divergence between sympatric hybrids and parentals. This contrasts with other well-known systems in which hybrids exhibit intermediate, novel, or transgressive phenotypes. Finally, the genetic basis of these phenotypes is unclear and further data are needed to clarify the evolutionary basis of morphological changes with ecological consequences.

evolutionary biology↗

bioRxiv2023

Unsupervised Machine Learning for Species Delimitation, Integrative Taxonomy, and Biodiversity Conservation

Integrative taxonomy combining data from multiple axes of biologically relevant variation is a major recent goal of systematics. Ideally, such taxonomies would be backed by similarly integrative species-delimitation analyses. Yet, most current methods rely solely or primarily on molecular data, with other layers often incorporated only in a post hoc qualitative or comparative manner. A major limitation is the difficulty of deriving and implementing quantitative parametric models linking different datasets in a unified ecological and evolutionary framework. Machine Learning methods offer flexibility in this arena by learning high-dimensional associations between observations (e.g., individual specimens) across a wide array of input features (e.g., genetics, geography, environment, and phenotype) to delineate statistical clusters. Here, I implement an unsupervised method using Self-Organizing (or "Kohonen") Maps (SOMs). Recent extensions called SuperSOMs can integrate an arbitrary number of layers, each of which exerts independent influence on the two-dimensional output clustering via empirically estimated weights. These output clusters can then be delimited into K significant units that are interpreted as species or other entities. I show an empirical example in Desmognathus salamanders with layers representing alleles, space, climate, and traits. Simulations reveal that the SOM/SuperSOM approach can detect K=1, does not over-split, reflects contributions from all layers with signal, and does not allow layer size (e.g., large genetic matrices) to overwhelm other datasets, desirable properties addressing major concerns from previous methods. Finally, I suggest that these and similar methods could integrate conservation-relevant layers such as population trends and human encroachment to delimit management units from an explicitly quantitative framework grounded in the ecology and evolution of species limits and boundaries.

evolutionary biology↗

bioRxiv2019

Phylogenetic and Spatial Distribution of Evolutionary Isolation and Threat in Turtles and Crocodilians (Non-Avian Archosauromorphs)

The origin of turtles and crocodiles and their easily recognized body forms dates to the Triassic. Despite their long-term success, extant species diversity is low, and endangerment is extremely high compared to other terrestrial vertebrate groups, with ~ 65% of ~25 crocodilian and ~360 turtle species now threatened by exploitation and habitat loss. Here, we combine available molecular and morphological evidence with machine learning algorithms to present a phylogenetically-informed, comprehensive assessment of diversification, threat status, and evolutionary distinctiveness of all extant species. In contrast to other terrestrial vertebrates and their own diversity in the fossil record, extant turtles and crocodilians have not experienced any mass extinctions or shifts in diversification rate, or any significant jumps in rates of body-size evolution over time. We predict threat for 114 as-yet unassessed or data-deficient species and identify a concentration of threatened crocodile and turtle species in South and Southeast Asia, western Africa, and the eastern Amazon. We find that unlike other terrestrial vertebrate groups, extinction risk increases with evolutionary distinctiveness: a disproportionate amount of phylogenetic diversity is concentrated in evolutionarily isolated, at-risk taxa, particularly those with small geographic ranges. Our findings highlight the important role of geographic determinants of extinction risk, particularly those resulting from anthropogenic habitat-disturbance, which affect species across body sizes and ecologies.

evolutionary biology↗