New DNA barcodes close a reference gap for South African sharks and rays

Environmental DNA metabarcoding promises to reveal which sharks, rays and chimaeras are swimming in a stretch of ocean without anyone having to catch or even see them. The method works by fishing fragments of DNA out of seawater and matching them against a library of known reference sequences. When a species is missing from that library, its DNA in the sample simply cannot be named. A new study reports a substantial expansion of the reference library for South African chondrichthyans, the class that includes sharks, rays and chimaeras, built around the mitochondrial 12S ribosomal RNA gene.

Researchers sequenced 69 individuals representing 36 chondrichthyan species from South Africa’s Indian and Atlantic Ocean coastlines, adding barcodes for species that previously lacked any 12S rRNA reference. The work lifted regional reference coverage from 50.3% to 69.1% of the country’s listed chondrichthyan species and added 18 species to the global pool of available 12S rRNA barcodes. Because 30 of the newly sequenced species also occur outside South African waters, the authors expect the new sequences to support eDNA studies far beyond the region where the tissue samples were collected.

The scale of the gap that motivated the study is striking. According to a GenBank search conducted on 30 April 2025, only about 17% of the estimated 1,250 to 1,300 chondrichthyan species worldwide had a 12S rRNA reference barcode available. In South African waters the picture was similarly incomplete: 56.7% of shark species, 43% of ray species and 62% of chimaera species had reference barcodes. That means roughly half of the region’s known elasmobranch species would be undetectable in an eDNA metabarcoding survey, and missing references can inflate the number of ambiguous assignments that mask true regional diversity.

To fill those gaps, the team sourced tissue samples for 41 South African chondrichthyan species that lacked 12S rRNA barcodes, drawing on archived specimens from an evolutionary genomics group, university collaborators, a national aquatic biodiversity institute and a shark conservation organization. Ninety-six tissue samples representing 41 species were amplified, although five species, including the sharpnose sevengill shark Oxynotus centrina and the puffadder shyshark Haploblepharus pictus, failed to amplify or did not yield usable sequences, leaving 36 species successfully added. The resulting sequences were uploaded to GenBank under accession numbers PX741113 through PX741183.

A central technical contribution of the study is a newly developed primer pair, V05F_898E and teleoR_Elas, designed to amplify a longer stretch of the 12S rRNA gene than the short fragments typically recovered from environmental samples. The researchers evaluated published fish primers in silico against 44 whole mitochondrial genomes of South African chondrichthyan species and found that several widely used sets lacked sufficient complementarity to local taxa. The selected primers were slightly modified to eliminate mismatches present across all tested sequences, and the final pair produced no mismatches with the tested genomes. The expected amplification products range from 391 to 702 base pairs, and the amplified region corresponds to positions 325 to 1198 of the reference mitochondrial genomes, overlapping both the 12S rRNA gene and the commonly used Elas02 eDNA barcode region.

That overlap matters for practical reasons. The Elas02 primer set, a taxon-specific modification of the widely used MiFish E primers, was developed precisely because general teleost primers showed amplification biases in biodiverse environments, and one study using the MiFish E primer set failed to detect any elasmobranch DNA along the entire South African coastline. The new barcodes, roughly 699 base pairs long, were deliberately designed to overlap the Elas02 region so that tissue-derived reference sequences can anchor the identification of shorter eDNA amplicons. The authors note that amplification of such long fragments from environmental DNA is generally constrained by DNA degradation, so the new primers are not intended for direct use in eDNA workflows; their role is to generate high-quality reference barcodes that improve taxonomic assignment of the shorter fragments actually recovered from seawater.

To verify the authenticity of the new sequences, the researchers compared them with GenBank entries and built Bayesian phylogenetic trees for batoids, selachians and holocephalans separately, using 249 sequences in total after combining the new barcodes with downloaded references. The analysis showed that 97.2% of the newly generated sequences clustered within their expected family-level lineages. Most recognized families formed distinct, well-supported clades. Twenty-one of 25 selachian families were monophyletic, and the holocephalan tree cleanly resolved all three recognized families, Callorhinchidae, Chimaeridae and Rhinochimaeridae, with no misplacements detected. Percent identity analyses reinforced the marker’s discriminatory power: intraspecific comparisons consistently exceeded 99.6% sequence identity, while interspecific comparisons averaged substantially lower values, and Wilcoxon rank-sum tests confirmed significant separation between the two categories for all three groups.

The study also documents where the marker falls short, and the authors are explicit that these limitations matter for interpretation. Six of 11 batoid families failed to show monophyly, with the newly sequenced Cruriraja hulleyi disrupting the monophyly of Rajidae, and four selachian families, including Carcharhinidae and Lamnidae, were not monophyletic. The porbeagle Lamna nasus did not group with other Lamnidae sequences despite a confirming BLAST match, a placement the authors attribute to the limited resolution of the mitochondrial fragment for deeper lamniform relationships rather than to misidentification. Within Carcharhinidae, minimal divergence separated congeneric species such as Carcharhinus brachyurus and Carcharhinus leucas, which exceeded 99% identity, a pattern consistent with prior reports of mitochondrial homogeneity, incomplete lineage sorting and hybridization in the genus. Similar high interspecific similarity between Mobula birostris and M. alfredi reflects known cryptic diversity among mobulid rays. The authors caution that species boundaries and family-level placement in batoids remain uncertain and that a single 12S rRNA reference database should be used with care, especially in taxonomically complex groups.

Comparative analyses quantified the cost of working with the short fragment that eDNA studies actually produce. When the phylogenetic analyses were repeated on alignments trimmed to the Elas02 length, roughly 216 to 243 base pairs depending on the group, phylogenetic resolution was significantly reduced. Internal branches shortened into star-like topologies in parts of the batoid tree, several Mobula species were represented by individual or weakly supported nodes, and pentanchid genera were scattered across the selachian tree rather than forming a cohesive clade. Percent identity comparisons showed the same pattern: intra- and interspecific values remained clearly separated in both datasets, but the magnitude of separation was reduced in the Elas02 dataset, with lower mean interspecific identity values and broader overlap between the categories, and interspecific distributions differed significantly between the short and full-length datasets. The authors conclude that the Elas02 region remains suitable for accurate amplicon sequence variant matching in eDNA metabarcoding, while the longer fragment generated here provides improved phylogenetic resolution that may help identify unknown or ambiguous ASVs.

The completed library now comprises 249 sequences from 131 species, including 102 batoid, 119 selachian and 28 holocephalan sequences, and the addition of new barcodes brought 21 families to complete species-level coverage. Reference sequences were added for nine batoid, 27 selachian and one holocephalan species, including the region’s only recorded angel shark species, Squatina africana, and the three newly added Scyliorhinidae species raised that family’s coverage to 100%. The remaining 30.9% of South African chondrichthyans without a 12S rRNA barcode are primarily deep-water species, rare taxa, or species with uncertain occurrence in the region, and the authors argue that natural history collections and museum specimens will be crucial for reaching them. The study does not assess ecological patterns or temporal changes in species distributions, but the authors position the improved coverage as a baseline for future work, noting that several sequenced species are of conservation concern, including the critically endangered tope shark Galeorhinus galeus. As eDNA metabarcoding becomes embedded in biodiversity monitoring and policy, the findings support a growing consensus that reference library completeness often matters more than marker choice in applied eDNA research.

Subject of Research: Expansion of the mitochondrial 12S rRNA reference database for South African sharks, rays and chimaeras to support environmental DNA metabarcoding

Article Title: Enhancing the 12S rRNA reference library of South African chondrichthyans to aid in eDNA metabarcoding

Article References: Loosen, K., Andreotti, S., & Matthee, C. A. (2026). Enhancing the 12S rRNA reference library of South African chondrichthyans to aid in eDNA metabarcoding. BMC Zoology, 11(1), Article 22. https://doi.org/10.1186/s40850-026-00272-4

Image Credits: AI Generated

DOI: 10.1186/s40850-026-00272-4

Keywords: eDNA metabarcoding, 12S rRNA, chondrichthyans, sharks, rays, chimaeras, reference library, South Africa, Elas02 primers, phylogenetics, GenBank, biodiversity monitoring