Taxonomic Turmoil: Using ddRAD Data to Resolve Taxonomic Discrepancies Within the Genus Characodon (Cyprinodontiformes: Goodeidae)
Characodon (Cyprinodontiformes: Goodeidae) is a highly imperiled genus of fishes from the Río Mezquital basin in northern Mexico that consists of two extant species (C. lateralis and C. audax), as well as a third extinct species (C. garmani), only known from the holotype. There are uncertainties in regard to the type localities for C. lateralis and C. garmani, making taxonomic revisions challenging. We gathered ddRADSeq data to generate a genome-wide dataset for populations of Characodon across the range. Our population genetic (FST, DAPC, PCA, sNMF) and phylogenetic (ML) analyses recovered three major groups/clades within Characodon, including one group corresponding to above the falls (Cascada El Saltito) and two separate groups from below the falls. Species delimitation analyses using BPP confirm the existence of three groups. Therefore, we recognize four species within Characodon. We must continue to recognize C. garmani as we did not include samples in our study and are therefore unable to re-assess its taxonomic status based on genetic data. We recognize all other populations above the falls in the Río Mezquital basin as C. lateralis and recognize C. audax as a junior synonym. Finally, we recognize two undescribed species below the falls, Characodon sp. 1 (Los Berros-La Constancia) and Characodon sp. 2 (Amado Nervo) that await formal descriptions.
Distribution of Characodon (Goodeidae) in northern Mexico. Red stars represent the location of the waterfall known as El Salto or Casacada El Saltito. Open circles represent archived museum records of Characodon (https://www.fishnet2.net, May 2024). Orange circles correspond to populations included in this study from the above the falls clade, and blue circles correspond to populations from the below the falls clade. The green circle in the offset map corresponds to the type locality of C. garmani.
Maximum likelihood phylogenetic analysis performed on a 15% complete sequence matrix in IQ-TREE. All nodes are recovered with bootstrap support values of 100 with the exception of the three nodes indicated by light gray dots. The scale bar represents branch lengths in units of substitutions per site per million years.
Principal component analysis (PCA) of unlinked single nucleotide polymorphism (SNP) data from the 75% complete data matrix. Each dot corresponds to an individual in our dataset and is colored according to population assignment. Ellipses indicate 95% confidence intervals. The inset in the bottom left-hand corner depicts eigenvalues representing variance captured by each principal component.
(A) Scatterplot visualization of the results of discriminant analysis of principal components (DAPC) under a clustering scenario of k = 3 genotypic clusters within Characodon. Points represent individuals and are colored according to assignment of individuals to clusters identified by the k-means algorithm. (B) Barplot visualizing the results of sNMF under clustering scenarios of k = 2–4. Vertical bars represent individuals and are colored according to the probability of membership of that individual to a particular cluster. In both panels, orange corresponds to the population above the falls, light blue represents Amado Nervo (below the falls), and dark blue represents Los Berros (below the falls).
Species delimitation analysis performed in the BPP program provides support for three distinct species of Characodon. The phylogenetic tree represents the best tree inferred using an IQ-TREE analysis of 200 ddRAD loci representing 30 individuals of Characodon. Heatmaps at each node reflect the posterior probabilities that the node represents a species-level bifurcation under each species delimitation algorithm and set of priors on tau and theta (see inset in top left-hand corner).
Contributor Notes
Associate Editor: R. E. Reis.