Echinoderms are marine water invertebrates that are represented by more than 7000 extant species, grouped in five classes and showing diverse morphologies (starfish, sea lilies, feather stars, sea urchins, sea cucumbers, brittle and basket stars). In an effort to further study their diversity, DNA barcodes (DNA fragments of the 5’ end of the cytochrome c oxidase subunit I gene, COI) have been used to complement morphological examination in identifying evolutionary lineages. Although divergent clusters of COI sequences were reported to generally match morphological species delineations, they also revealed some discrepancies, suggesting overlooked species, ecophenotypic variation or multiple COI lineages within one species. Here, we sequenced COI fragments of 312 shallow-water echinoderms of the East Coast of South Africa (KwaZulu-Natal Province) and compared morphological identifications with species delimitations obtained with four methods that are exclusively based on COI sequences. We identified a total of 103 morphospecies including 18 that did not exactly match described species. We also report 46 COI sequences that showed large divergences (>5% p-distances) with those available to date and publish the first COI sequences for 30 species. Our analyses also identified discordances between morphological identifications and COI-based species delimitations for a considerable proportion of the morphospecies studied here (49/103). For most of them, further investigation is necessary to keep a sound connection between taxonomy and the growing importance of DNA-based research.
Citation: Sonet G, Smitz N, Vangestel C, Samyn Y (2022) DNA barcoding echinoderms from the East Coast of South Africa. The challenge to maintain DNA data connected with taxonomy. PLoS ONE 17(10): e0270321. https://doi.org/10.1371/journal.pone.0270321
Editor: Tzen-Yuh Chiang, National Cheng Kung University, TAIWAN
Received: June 7, 2022; Accepted: September 21, 2022; Published: October 10, 2022
Copyright: © 2022 Sonet et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All DNA sequences are available in the Barcode of Life Data Systems (BOLD) with Process IDs from CHARE001-20 to CHARE400-20 (project CHARE).
Funding: The DNA analysis performed in this study was financed by the Belgian Science Policy. Funding for this work also came from the Flemish Community Bilateral (International) Scientific and Technological Cooperation, project numbers BIL98/84 (YS) and BIL01/46 (YS), the Fund for Scientific Research Flanders (YS), the Research Council of the Free University Brussels (YS), the Belgian Global Taxonomic Initiative (YS) and the King Leopold III Fund for Nature Exploration and Conservation (YS). These funders had no role in study design, data collection and analysis, decision to publish, or manuscript redaction.
Competing interests: The authors have declared that no competing interests exist.
Echinoderms are exclusively marine water invertebrates whose larvae are bilaterally symmetric while the adults show the typical pentamerous radial symmetry, which can be doubled by a secondary bilateral symmetry (cf. holothuroids or irregular echinoids). Their phylum (Echinodermata) includes five extant classes: Asteroidea (starfish or sea stars), Crinoidea (sea lilies and feather stars), Echinoidea (sea urchins), Holothuroidea (sea cucumbers) and Ophiuroidea (brittle and basket stars). They are represented by more than 7000 extant and 13000 extinct species  but this number likely is a gross underestimation given that taxa continue to be described annually [2–7]. In addition to morphology, a mitochondrial DNA region of the 5’ end of the cytochrome c oxidase subunit I gene (COI) is widely used as a DNA barcode for the identification of animal species . It has been used as a useful proxy to study the diversity of extant echinoderm species because clusters of similar COI sequences largely corroborated morphological identifications in several large-scale studies including all five classes of echinoderms from Canadian waters , the coasts of Australia and New Zealand  and the North Sea . Yet, divergent clusters of COI sequences were observed within some species [9–12]. Some of these divergent clusters were assigned to undescribed species on the basis of detailed taxonomic analyses combining morphological characteristics, DNA data and sometimes ecological observation [13–17]. Divergent clusters of COI sequences were also observed within Linckia multifora (distances up to 0.023) and L. laevigata (divergent mitogenomes) [18, 19], but could not be associated with morphological differentiation, suggesting either cryptic speciation or hybridization. Similarly, an integrative taxonomical study on the ophiuroid Astrotoma agassizii suspected that seven well-supported and deeply divergent COI clades found in this group corresponded to cryptic species despite evidence of admixture [20, 21]. In contrast, divergent COI lineages (0.011–0.044 distances) were shown to coexist in two species of the Ophioderma longicauda species complex recently delimited based on morphology, developmental features, and mitochondrial and nuclear (coding and non-coding) DNA [22, 23]. These studies show that clusters of COI sequences can reveal lacunes of knowledge in echinoderm species diversity and that integrative taxonomy brings added value to address some species complexes.
In the Republic of South Africa, despite a relatively strong tradition of marine taxonomic research  and the availability of voucher specimens from historical and recent biodiversity inventories, the shallow-water echinoderms of the North-East coast of the country have until recently remained poorly documented. In addition, there is a poor representation of DNA barcodes from South African echinoderms in the online DNA repositories compared to other regions of the world. In fact, less than 500 public COI sequences are available for South African echinoderms out of more than 34000 public sequences in the Public Data Portal of BOLD . In this context, we undertook the DNA barcoding of recently collected shallow-water echinoderms of the North-East coast of South Africa (KwaZulu-Natal Province). Our goals are to (i) provide DNA barcodes for recently (1999–2016) sampled shallow-water echinoderms from the North and South of the KwaZulu-Natal coast, (ii) compare morphological identifications with species delimitations exclusively based on COI sequences and (iii) flag specimens for which additional taxonomic investigations are needed. This exercise is useful to describe the species diversity of echinoderms in South Africa but also on earth, as South Africa is situated in the West of the Indian Ocean, which is known for its exceptional marine diversity .
Materials and methods
Sampling and morphospecies identification
A total of 312 specimens were collected during five different campaigns (from 1999 to 2016) in the shallow-waters North and South of the KwaZulu-Natal Province, on the East Coast of the Republic of South Africa (Fig 1): 41 asteroid, 90 crinoid, 23 echinoid, 70 holothuroid and 88 ophiuroid specimens. All specimens studied here are invertebrates and were collected legally with written permissions from the regional and national South African authorities. The permit number of the expedition of 2016 is RES 2016–02 and was granted by the Department of Agriculture, Forestry and Fisheries (DAFF) and the Department of Environmental Affairs (DEA) of South Africa to Yves Samyn. The specimens were collected in the eulittoral zone whilst snorkelling or in the sublittoral zone to ca. 40 m depth while scuba diving. Specimens were anesthetized with ± 5% Mg2.6H20 (or freshwater for crinoids and some ophiuroids), fixed with 80–95% ethanol for at least 24 hours, rinsed with 70–75% ethanol and finally stored either in 70–75% ethanol or dry. Tissue samples, generally dermis, podia or tips of arms were sampled in the field and stored in absolute ethanol for subsequent DNA analysis. They were deposited in the Royal Belgian Institute of Natural Sciences (Belgium), the Royal Museum for Central Africa (Belgium) and the University of KwaZulu-Natal and the Iziko Museums of South Africa (Republic of South Africa). Voucher numbers and specimen information are provided in S1 Table. Taxonomical identifications were performed on the bases of external and internal anatomy and internal skeletal elements. For this, specimens were either denuded from their dermis to reveal the underlying calcareous skeleton (for instance test and spine structure of echinoids) or dissected to isolate various (microscopic) calcareous deposits (for instance the ossicle assemblage of holothuroids). Identifications were achieved by eye, stereo light microscope or scanning electron microscope. General identification keys [27, 28] were complemented with more detailed taxon- and region-specific revisions as well as with more recent taxonomic literature. Taxonomy is following the World Register of Marine Species . When possible, specimens were compared with vouchers stored in museum collections. These identifications are hereunder designated as “morphospecies”. In some cases, specimens could not be identified to the species level, either because diagnostic taxonomic characters could not be assessed or because of differences between the specimen and species descriptions. In these cases, provisional identifications were given the taxonomic rank with the highest confidence (e.g. Ophiocoma sp. or Amphiuridae sp.).
A. Map of the East Coast of South Africa with sampling locations represented as black dots, country names in bold, and sampling location names in normal font. B. Map of Africa where the zoomed geographic region is represented as a white rectangle outline. This map was created using the software QGIS v. 3.22.6 , the GEBCO Grid  and Natural Earth .
DNA data collection
Total genomic DNA was extracted using the NucleoSpin Tissue Kit (Macherey-Nagel, Germany) following the standard protocol for animal tissue. Fragments of the 5’ end of the cytochrome c oxidase subunit I (COI) gene were amplified and sequenced using the following pairs of primers (sequence provided from 5’ to 3’): LCO1490 (GGTCAACAAATCATAAAGATATTGG) with HCO2198 (TAAACTTCAGGGTGACCAAAAAATCA) , EchinoF1 (TTTCAACTAATCATAAGGACATTGG) with EchinoR1 (CTTCAGGGTGTCCAAAAAATCA) , COIeF (ATAATGATAGGAGGRTTTGG) with COIeR (GCTCGTGTRTCTACRTCCAT)  and LoboF1 (KBTCHACAAAYCAYAARGAYATHGG) with LoboR1 (TAAACYTCWGGRTGWCCRAARAAYCA) . When these primer pairs were not successful, other combinations of forward and reverse primers were tested and provided amplicons (LCO1490/EchinoR1, LCO1490/COIeR, EchinoF1/HCO2198, EchinoF1/COIeR COIeF/HCO2198 and COIeF/EchinoR1). The polymerase chain reactions (PCR) were prepared in volumes of 25 μl containing 2 μl of DNA template, 0.03 U/μl of Platinum® Taq DNA Polymerase (Life Technologies, USA), 1X PCR buffer, 0.2 mM dNTPs, 0.4 (all primer pairs except LoboF1/LoboR1) or 0.6 (LoboF1/LoboR1) μM of each primer and 1.5 mM MgCl2. The PCR profiles started with one step at 94°C for 3 min. Then, for all primer pairs except LoboF1/LoboR1, the next step consisted of 40 cycles at three temperatures: 94°C for 30 s, 48°C for 30 s and 72°C for 45 s. For the primer pair LoboF1/LoboR1, five first cycles were performed at 94°C for 30 s, 45°C for 90 s and 72°C for 60 s, before 40 additional cycles at 94°C for 30 s, 54°C for 90 s and 72°C for 60 s. All PCR profiles ended with a final step at 72°C for 7 min. PCR products were visualized using 1.2% agarose gel electrophoresis and purified using the ExoSAP procedure (Exonuclease I—Shrimp Alkaline Phosphatase from ThermoFisher, USA). PCR products were sequenced bi-directionally on an ABI automated capillary sequencer (ABI3130xl) using the BigDye v3.1 chemistry following the manufacturer’s instructions (Life Technologies, USA). DNA chromatograms were checked, trimmed and assembled using CodonCode Aligner© v8.0.2 (CodonCode Corp., Centerville, Massachusetts). Consensus sequences were inspected (detection of gaps and stop codons) and compared to public records using the Basic Local Alignment Search Tool (BLAST)  of the National Centre for Biotechnology Information, U.S. National Library of Medicine (NCBI) to detect and exclude obvious contaminations.
DNA barcode analysis
The COI sequences generated here were compared with those available in the Barcode of Life Data System (BOLD) . For this, the “advanced search of records” was used from the workbench, whereby the name of each class was searched in the field “taxonomy” and both the target marker “COI-5P” and the option “include public records” were selected. All retrieved records were downloaded after applying the built-in filters to exclude the sequences with stop codons and tagged as contaminants. For each class, a dataset was assembled, containing both the public sequences (below called “BOLD sequences”) and those obtained here, aligned using the MUSCLE algorithm  on MEGA v7.0.26 . Shorter sequences were removed to keep at least 150 bp common to all aligned sequences. Proportions of differences (uncorrected p-distances) among all sequences in each alignment were calculated using the package ape  in the R language and environment for statistical computing and graphics. The “Search Taxonomy” tool of BOLD  was used on the 24 May 2022 to check the availability of public COI sequences for each species identified here. For specimens identified with provisional species names, the identification engine of BOLD was used to search for best matches in the “All Barcode Records” database on BOLD, which includes COI sequences that have not yet been released publicly (“private records”). The Process IDs of the DNA sequences in the Barcode of Life Data Systems (BOLD) are from CHARE001-20 to CHARE400-20 (project CHARE).
COI-based species delimitation
We applied four methods to delimit putative species on the basis of DNA barcodes and without a priori knowledge of species identity: (i) the automatic barcode gap discovery (ABGD) method , (ii) the “Barcode Index Numbers” (BIN) based on Refined Single Linkage (RESL) analysis , (iii) the General Mixed Yule Coalescent (GMYC) model  and (iv) a Bayesian implementation of the Poisson tree processes (bPTP) model . These delineations of putative species based on DNA barcodes represent a repeatable way to recognize operational taxonomic units (OTUs), which can be used with other lines of evidence for species assignment . The four methods were applied to each class separately. The first two methods are based on distances among COI sequences and the two others are based on the phylogenetic species concept for which an inference of phylogeny was performed (see below). For the ABGD method, the web application  was used to enable the automatic detection of gaps in the distribution of pairwise distances among DNA barcodes, which can be used to delimit putative species. The BINs were obtained on BOLD., BINs are regarded as proxies to taxonomic species. These BINS result from a single linkage clustering of the sequences, which were then grouped or split based on their similarity and connectivity. When the sequence overlap with the BOLD sequences was lower than 300 bp and a BIN was not attributed directly by BOLD, the BIN of the best match above 99% similarity was assigned to the record. For the GMYC method, we used the R package “splits” with the single threshold option  to determine the transition point between inter- and intra-species branching rates on a time-calibrated ultrametric tree  constructed as detailed in the section below. In the bPTP model, the transition points between inter- and intra-species branching are estimated from mean expected number of substitutions per site between two branching events . Congruence among the species identifications based on morphology and on COI sequences was analysed and discordances were classified in three categories: (R) “resolution issues” when representatives of the same morphospecies were subdivided in several DNA-based putative species by the different species delimitation methods; (S) “splitting issues” when these representatives were split in several DNA-based putative species by all species delimitation methods and (L) “lumping issues” when different morphospecies were grouped in one putative species by all DNA-based species identification methods. We also tagged the discordance if it was observed only within the data produced here (H) or only after merging our data with all data (A) available online.
Phylogenetic trees were reconstructed through Bayesian inference (BI) and were used as an input for both GMYC and bPTP. The program MrBayes v3.2.7a  was used with the strict clock model . The analysis was run on the CIPRES Science Gateway  using the best partition scheme and best-fit substitution models estimated using PartitionFinder v. 1.1 . As datasets included 1384–4898 sequences, we saved computation time by performing the BI with representative unique haplotypes collected as follows: the ABGD approach was applied to the whole dataset using parameters that tend to oversplit the samples (prior maximal p-distance (P) ranging from 0.0001 to 0.1 and a relative gap width of 1). The oversplit resulting from the partition with a P = 0.0001 was checked in the graph representing the number of groups obtained for the selected range of prior intraspecific divergences. Three sequences were then collected randomly from each of these groups (less when the group consisted in one or two sequences). Unique haplotypes were then extracted from this selection using the R language and environment and the pegas package . For BI, two parallel runs were run for 50 million generations. Convergence was checked and the first 25% of the tree sampled were discarded as “burn-in” in order to discard the preliminary steps, during which the chain moves from its unrepresentative initial value to the modal region of the posterior . The majority-rule consensus trees reconstructed here are available as S1–S5 Files.
Based on morphology, the 312 echinoderm specimens sampled in this study were assigned to 103 morphospecies belonging to the five extant classes of Echinodermata (Tables 1 and 2). Of these 103 morphospecies, 18 could not be assigned to any described species and were given a provisional name (Tables 1 and 2). The 312 DNA sequences generated here were fragments of 420 to 841 base pairs (bp) (arithmetic mean of 692 bp) of the 5’ end of the cytochrome c oxidase subunit I gene (COI). For the vast majority of the morphospecies (91/103), COI sequences showed at least one substitution compared with the sequences available so far in public repositories. For 46 morphospecies, COI sequences even showed more than 5% p-distance difference with their best match in public repositories (Fig 2). For 30 of the 85 morphospecies that could be assigned to a described species, we publish here the first COI sequences (Table 2). The sequences were grouped by the four species delimitation methods in 99 (ABGD), 103 (GMYC and bPTP) and 105 (BIN) putative species (Table 2).
The p-distances (proportion of sites at which two sequences are different) separating each DNA sequence obtained here and its best matches in BOLD are plotted (maximum 100 best matches with a Process ID, no provisional names and a minimum of 80% similarity were considered). The p-distances are grouped for each morphospecies (x-axis) and plotted as green crosses when species names matched, as red circles when species names did not match or as black triangles when morphospecies with provisional names were included in the comparison. Distances >20% were measured between conspecifics of Echinoneus cyclostomus, Holothuria difficilis, Holothuria impatiens and Ophiothrix fragilis.
Among the 41 starfishes sequenced here, 18 different morphospecies were identified: 15 were recognized as species based on the current taxonomy and three were identified with provisional names (Aquilonastra cf. rowleyi, Aquilonastra sp. and Leiaster cf. leachi). Species delimitation methods assigned the 41 COI sequences to a total of 18 (ABGD), 20 (GMYC and bPTP) or 21 (BIN) putative species (Table 1). Discordances between morphospecies assignments and DNA-based identification methods are detailed in Table 3. The number of putative species determined with GMYC, bPTP and BIN was higher than with ABGD because Fromia milleporella was subdivided in two putative species by all methods except ABGD, and Gomophia egyptiaca was lumped with Nardoa variolata in one putative species by ABGD only. BIN was the only method to split the sequences obtained for Aquilonastra sp. in two putative species. When also considering BOLD data, we note that the sequences of four species were subdivided in two or more putative species (Table 3): Our sequences of Fromia indica, Ophidiaster hemprichi and Echinaster purpureus were in a putative species separated from those from the Western Pacific Ocean or the coasts of Saudi Arabia. For Linckia laevigata only a few BOLD sequences were separated from conspecifics (including ours), but they did not seem to have a different geographic origin. Sequences of Linckia laevigata were also lumped with all BOLD sequences of Linckia multifora by all methods. The three morphospecies with provisional species names represented three (ABGD, GMYC and bPTP) to four (BIN) putative species that were separate from any other species for which COI sequences were available (Table 4). Indeed, Aquilonastra sp. represented one or two separate putative species (see above). Leiaster cf. leachi also formed a separate putative species, with Leiaster glaber Peters, 1852 (distance of 10.34%) as the closest published record. We noted a closer match (1.08–1.68%) with two private BOLD sequences labelled as L. leachi using the identification engine of BOLD. Finally, the specimens identified as Aquilonastra cf. rowleyi showed COI sequences that were split in two putative species, one of which was separate from any other putative species, the other was grouped with BOLD records also labelled with a provisional name (Aquilonastra sp.2).
The 90 crinoids analysed here were assigned to eight morphospecies, all of them corresponding to described species except one, Comanthus sp. Our crinoid specimens were grouped into eight (ABGD) or ten (GMYC, bPTP and BIN) putative species by the species delimitation methods (Table 1) because Oligometra serripinna and Tropiometra carinata were both subdivided in two putative species by the latter three methods. The specimens of Oligometra serripinna sequenced here and split in two putative species were both found in the iSimangaliso Wetland Park, not far from Sodwana Bay (Fig 1A). One of these putative species also included BOLD records from Singapore. For Tropiometra carinata, specimens of one putative species were collected in Protea Banks while specimens from the other putative species were from Umkomaas, Manzengwenya and Sodwana Bay (Fig 1). Compared to the data available in BOLD, discordances were noted for four further species (Table 3). Indeed, our COI sequences of Cenometra emendatrix were lumped with BOLD records of C. bella. For Comanthus wahlbergii, all methods split our sequences from the other sequences available online for the same species. For Dichrometra palmata, the sequences obtained from South Africa were split from one BOLD sequence originating from the East coast of Australia by GMYC, bPTP and BIN. Finally, the same three methods also split Stephanometra indica in two putative species, one of which included our sequence and three BOLD sequences from Australia and Papua New Guinea. The other was represented by one single sequence from the West Indian Ocean. The sequence obtained from the single specimens with a provisional name (Comanthus sp.) was either separated from all other putative species (ABGD, bPTP and GMYC) or grouped with Comanthus suavia and C. parvicirrus (ABGD), with no better match found by the BOLD identification engine among private data (Table 4).
The 23 echinoids studied here represented 12 morphospecies, three of which were given provisional identifications: Diadema sp., Irregularia sp. and Temnopleurus cf. toreumaticus. COI-based species delimitation methods identified 11 (ABGD, GMYC and bPTP) or 12 (BIN) putative species as BIN was the only method that did not lump Diadema savignyi with Diadema sp. (Table 3). Concerning the discordances with BOLD records, the same sequences of Diadema savignyi and Diadema sp. were also lumped by ABGD, GMYC and bPTP with BOLD sequences of D. mexicanum, D. antillarum and Echinothrix diadema. Similarly, sequences of Echinometra mathaei were either lumped with Echinometra oblonga (ABGD and GMYC) or split in minimum three putative species (bPTP and BIN). Our COI sequences of Echinostrephus molaris were lumped by ABGD and GMYC in the same putative species as one BOLD record identified as Heterocentrotus trigonarius, but which was tagged as misidentified. BIN was the only method to subdivide Eucidaris metularia in three putative species, with one putative species only represented by specimens from Hawaii, and two putative species including both specimens from Saudi Arabia (Red Sea) and South Africa. Tripneustes gratilla was grouped by all methods with Tripneustes depressus and a few sequences of Tripneustes kermadecensis. The COI sequences with a provisional identification (other than Diadema sp. already detailed above), Irregularia sp. and Temnopleurus cf. toreumaticus, were not grouped with any other COI sequence and did not match any private sequence in the identification engine of BOLD (Table 4).
Among the 70 sea cucumbers included in this study, 32 morphospecies were identified. Six of them were attributed provisional names (Bohadschia sp., 4 Holothuria sp. and Thyone sp.). Species delimitation methods distinguished 29 (BIN) or 31 (ABGD, GMYC and bPTP) putative species. The discordances between the species identified morphologically and using COI were caused by ten sequences of eight morphospecies that were mixed in three putative species by all methods. A first putative species grouped Holothuria albofusca with H. hilla; a second putative species grouped H. insignis, H. lineata, H. pardalis, H. tuberculata and H. sp. (voucher RMCA.2121), and the third putative species grouped H. hilla, H. impatiens, H. insignis and H. pardalis (Table 3). Conversely, several of these morphospecies were split in highly divergent putative species by all methods. Large distances were found within our sampling for Holothuria hilla (p-distance up to 19.74%), H. impatiens (16.45%), H. insignis (13.16%) and H. pardalis (15.79%). Large distances were also recorded for two of these species within the public data set available in BOLD (Holothuria hilla and H. impatiens with p-distances of 11.84–16.45% and 17.76–24.34%, respectively). Finally, when considering both BOLD and our dataset, additional species were split in distantly related putative species: Holothuria difficilis (up to 22.37%), Actinopyga mauritiana (15.79%), Holothuria atra (11.18%), Actinopyga obesa (7.89%) and Holothuria cinerascens (5.92%). One of the six specimens with a provisional identification (Thyone sp.), represented a putative species that was separate from any other putative species identified in BOLD or within our dataset, and had no match using the identification tool of BOLD. The sequence of two other specimens labelled as Holothuria sp. (RMCA.2129 and RMCA.2103) did not group with any other putative species but matched two private records of BOLD of Holothuria hartmeyeri and Holothuria rigida with dissimilarities of 0.50 and 0.77%, respectively. The three other specimens with a provisional identification were grouped with other morphospecies by all species delimitation methods: Bohadschia sp. was grouped with public sequences of B. cousteaui (BOLD Process IDs GBEHH226-13 and GBEHH227-13), Holothuria sp. RMCA.2634 was grouped with public sequences of H. isuga (BOLD Process ID GBEHH114-10) and Holothuria sp. RMCA.2121 was grouped with the putative species including different morphospecies mentioned above (Holothuria insignis, H. lineata, H. pardalis and H. tuberculata). This cluster also included two public records from unidentified holothuroids from Queensland, Australia (BOLD Process IDs ECLI026-08 and ECLI027-08). Using the identification tool of BOLD, two records of Holothuria arenacava showed a dissimilarity of 0.24–0.77% with Holothuria sp. RMCA.2634 (Table 4).
Among the 88 ophiuroids sampled here, a total of 33 morphospecies were identified. Five of these morphospecies were identified with a provisional name: Amphiuridae sp., Gorgonocephalidae sp., Ophiactis sp., Ophiocoma sp. and Ophiothela sp. Species delimitation methods distinguished 31 (ABGD, GMYC and bPTP) or 33 (BIN) putative species. This estimation difference is due to five COI sequences of Ophiactis picteti clustering in one single putative species according to ABGD, GMYC and bPTP and in three putative species according to BIN (Table 3). Remarkably, 15 ophiuroid morphospecies represented both in our sampling and in BOLD are split by all species delimitation methods in two or more putative species: Astroboa nuda (separated by p-distances up to 18.72%), Macrophiothrix longipeda (6.67%), Macrophiothrix propinqua (16.67%), Ophiactis savignyi (7.91%), Ophiarachnella gorgonia (10.40%), Ophiocoma pusilla (17.51%), Ophiocoma brevipes (10.98%), Ophiolepis cincta (16.56%), Ophionereis porrecta (18.11%), Ophiopeza fallax (12.69%), Ophiopeza spinosa (16.97%), Ophioplocus imbricatus (12.27%), Ophiothrix fragilis (23.85%), Ophiothrix purpurea (8.13% to 23.24%) if we consider a specimen that might be misidentified, see below) and Ophiothrix trilineata (10.37%). One additional morphospecies, Ophiocoma erinaceus (5.52%) was also split by GMYC, bPTP and BIN. Concerning the specimens with provisional identifications, the specimen labelled as Ophiactis sp. formed a putative species that was separate from all other putative species. The closest match in the BOLD identification engine was Amphipholis squamata with distances of 1.68–7.95. Of the two specimens identified as Ophiothela sp., one was grouped with four representatives of Ophiothela venusta and the other was grouped with six representatives of Ophiothrix purpurea. The specimens labelled as Amphiuridae sp.1, Gorgonocephalidae sp. and Ophiocoma sp. were grouped by all delimitation methods with Ophiocoma cf. doederleini (BOLD record), Astroboa nuda (sequenced here), and Ophiocoma pusilla (sequenced here), respectively (Table 4).
Further taxonomic investigation required for 48 morphospecies
DNA barcoding studies represent a complementary approach to enlarge our knowledge of species biodiversity. Associating short DNA sequences to morphospecies improves the characterization of evolutionary lineages and enables the detection of both erroneous identifications, synonyms and potentially undescribed species when discordances are observed between morphospecies and COI-based putative species identifications . Since COI alone is not suitable for delimiting and describing new species, the discordance observed require additional studies to redefine species on the bases of rigorous taxon delimitation methods and publicly available data [54, 55]. The total number of species identified based on morphology was overall consistent with DNA-based species delimitations (103 morphospecies versus 99–105 species estimated using COI). However, our results categorised 49 discordances (cf. “resolution”, “splitting” and “lumping” issues in Table 3 and discussed below) that can be explained by differences among COI-based species identification methods, imperfect species delineation (and misidentifications) and overlooked species diversity (or species complexes). Also, the morphospecies not matching described species can either represent unknown species morphs or undescribed species. These alternative interpretations are discussed below. In total, we identified at least 48 morphospecies (including eight with a provisional name) that could benefit from additional integrative taxonomical investigations to accurately identify them to the species level.
Discordances among COI-based species delimitation methods
All four DNA-based species delimitation methods provided congruent results for 81 of the 103 morphospecies. Part of the methods lumped four pairs of morphospecies in four putative species and split 14 morphospecies (cf. “R” issues in Table 3). Tree-based methods (GMYC, and bPTP) tend to overestimate the number of putative species while ABGD tends to underestimate them . For example, Ophiocoma erinaceus was split in two putative species by all methods except ABGD. Recent phylogenomic data based on exon capture suggested that the two putative species found in the West Indian and in the East Indian/Pacific Oceans underwent allopatric speciation . In the case of Fromia indica, the two different putative species identified by GMYC, bPTP and BIN were collected in different areas (South-western Indian Ocean and the Coral Sea in the South Pacific) and further investigation is necessary to know if they represent different species or divergent populations. For O. serripinna and T. carinata, also split in several putative species by all methods but ABGD, COI distances were not consistent with geographic distances, as separate putative species were collected in the same regions. A previous study including T. carinata from the Atlantic and Indian Oceans already revealed two divergent sympatric COI lineages, which likely represent different species . Only one of these two lineages (lineage 1) included our COI sequences (from Umkomaas, Manzengwenya and Sodwana Bay). The COI sequences obtained from the Protea Banks were forming a third putative species according to GMYC, bPTP and BIN and represents another case to be further investigated. For S. indica, the second putative species was represented by a single BOLD sequence (GBEH3087-10) that contained four gaps, and which should be double-checked.
BIN was shown to count more putative species than GMYC and bPTP . In our results, BIN was the only method in which the following morphospecies were split in two or more putative species. Sequences of Eucidaris metularia from the Red Sea and the Indian Ocean were split from those from the Pacific Ocean. Sequences of Holothuria edulis from the Red Sea and South Africa were split from others from the Gulf of Aqaba and the more eastern part of the Indo-Pacific Ocean. The five sequences of Ophiactis picteti sequenced here and originating from the same locality were split in three putative species. These last seven cases, but especially the O. picteti case, need to be further investigated. BIN was also the only method to distinguish Diadema savignyi, D. mexicanum and D. antillarum, three species that hybridize and that are known to be very closely related [60, 61]. BIN and bPTP were the only methods to split Echinometra mathaei in three or more putative species. This is in accordance with the distinct species identified within Echinometra mathaei using ecological distribution, test morphology, gonadal spicules, gametes and cross-fertilization experiments [62–66]. These cases show the difficulty to interpret the different species delimitation methods in absence of additional lines of evidence and when distances among species are relatively small. Putative species are considered more reliable when supported by various delimitation methods than when supported by only one . Therefore, the following discussion will focus only on putative species supported by all methods.
Imperfect species delineation
A total of 10 morphospecies (one asteroid, one crinoid, one echinoid and seven holothuroids, Table 2) showed COI sequences that were lumped with other morphospecies. Some were also split in different putative species. Many of these cases have already been reported in the literature. In asteroids, Linckia laevigata and Linckia multifora have been extensively sequenced, with hundreds of COI sequences and a geographically comprehensive sampling in the Indo-Pacific. Both species, although morphologically recognizable based on the skeletal armature , are hardly distinguishable using COI sequences , and may both comprise several cryptic species [18, 19, 68]. Concerning the crinoid morphospecies Cenometra emendatrix (known from the eastern Indian Ocean), our sequences were lumped with the BOLD sequences of Cenometra bella (known from the central to western Indo-Pacific) in one putative species. COI sequences of the two species were separated by small p-distances (0.92–1.72%), but one COI sequence of C. bella from Papua New Guinea (GBMND44777-21, voucher MNHN-342) was identical to one C. emendatrix from South Africa. Further analysis will decipher if C. emendatrix and C. bella result from a recent allopatric speciation or represent isolated populations from one single species. In echinoids, the lumping of Tripneustes gratilla (from the Indian and Pacific oceans) with Tripneustes depressus (from the eastern Pacific) in one putative species is consistent with a previous study based on morphology, COI and bindin gene data , which suggested that they belong to the same species. In holothuroids, the seven problematic morphospecies identified in this study are known to be difficult to distinguish using morphological characters. For example, Holothuria insignis and H. lineata have been confirmed as valid species based upon detailed studies of the recovered type material and one newly collected voucher specimen [70, 71]. These are difficult identifications, and species belonging to the same subgenera were often mixed in some putative species (Holothuria insignis, H. lineata, H. pardalis and H. tuberculata within the subgenus Lessonothuria Deichmann, 1958 or H. hilla and H. albofusca within the subgenus Mertensiothuria Deichmann, 1958). More importantly, several of these species were both split and mixed in several divergent COI-based putative species (likely representing a series of species complexes that cannot yet be properly delineated through morphological examination (in particular the microscopic ossicle assemblage) or that were erroneously identified. Specific focus on these species already revealed that one of the COI lineages represented by H. hilla has been recently described as a new species (Holothuria viridiaurantia sp. nov.) . Also, based on a combination of mitochondrial and nuclear markers, Michonneau  concluded that H. impatiens was a species complex comprising at least 13 species. Additional investigation including further comparisons with type specimens will be necessary to reliably associate at least nine of the 10 morphospecies discussed here based on DNA data, improve species delineation and resolve sometimes extensive subjective synonymy.
Overlooked species diversity
Most discordances between morphospecies and COI-based species delimitations were splitting issues (Table 2). Indeed, the COI sequences of 22 morphospecies (one asteroid, two crinoids, four holothuroids and 15 ophiuroids) were split in two or more putative species (Table 3) showing considerable divergences (5.92–23.85%) and deserving additional scrutiny. These cases can include both undescribed species and species already described but unrecognized morphologically and for which COI sequences are not yet available or in which divergent COI sequences coexist. For the holothuroid species Actinopyga mauritiana, our sequences and others from Egypt (Gulf of Aqaba) labelled as Actinopyga sp. nov. MA-2010  showed the highest similarity with A. varians (4.60%) from Hawaii and Micronesia, which is considered as its sister-species based on morphology . The other BOLD sequences from the Red Sea, also labelled as A. mauritiana could correspond to another species. Concerning A. obesa, the BOLD sequence from Hawaii that is separated from our sequences is grouped with A. caerulea, a species that was recently recognized and described . It is possible that A. caerulea was not yet known by the identifier who assigned another species name to the BOLD sequence from Hawaii. For Holothuria cinerascens, the separation of one sequence from India from those from South-Africa, Pakistan, Hawaii and China suggests that Holothuria cinerascens is largely distributed in the Indo-Pacific and that a distinct species was collected in India. Many of the 15 ophiuroid morphospecies showing a splitting issue were already reported to be composed of several putative species [77–79]. Ophiuroid species diversity is high in the Indo-Pacific  and was estimated to be underestimated by 20% in the South-western Indian Ocean . It is striking that the COI sequences obtained here for four ophiuroid species represented yet other putative species that were not yet sequenced. Our sequences of Astroboa nuda, Ophiopeza spinosa and Ophiothrix purpurea from South Africa were split from those from Australia and Japan. The sequence of Ophiothrix fragilis from South Africa was separated from the other O. fragilis sequences originating from the Atlantic and the Mediterranean Sea and forming three other COI-based putative species in our analysis. These three putative species were characterized by different geographic and bathymetric distributions [79, 81]. This underestimated species diversity in Ophiuroidea can explain why intraspecific COI distances measured for the currently described ophiuroid species were in average higher than for the other classes of echinoderms .
The varying levels of DNA sequence divergences observed within species may depend on sampling (representativity and geographic scale), life history and evolutionary history of the species. Species with lecithotrophic larvae such as Ophiarachnella gorgonia, Ophiolepis cincta, Ophionereis porrecta or Ophiopeza fallax are suspected to show a more limited dispersal capacity than species with planktotrophic larvae because they do not feed, and their development time is limited by their energy reserves. Therefore, species with lecithotrophic larvae often display COI with relatively larger intraspecific distances than species with planktotrophic larvae, enhancing the likeliness of allopatric speciation [9, 77, 78]. Based on these theoretical considerations, the existence of undescribed species was suspected when deep divergences were found within morphospecies with planktotrophic larvae . Our results show that the South-African representatives of the four above-mentioned species with lecithotrophic larvae were indeed grouped in the same putative species as the specimens sequenced from the South-western Indian Ocean but separated from the specimens sequenced from Australia or the Pacific Ocean. Yet, this pattern was also observed for other species with planktotrophic larvae (Macrophiothrix longipeda, Macrophiothrix propinqua, Ophiocoma brevipes, Ophiocoma pusilla, Ophioplocus imbricatus and Ophiothrix trilineata). These widespread species were suspected to be complexes of allopatrically-differentiated regional endemics by Boissin et al. , who also noticed morphological differences among the COI lineages of Ophiarachnella cf. gorgonia, Ophionereis porrecta and Ophiolepis cincta. However, a more comprehensive sampling might unveil haplotypes and reveal that the currently observed splits are artefacts resulting from limited sampling. Similar misleading observations were corrected for Linckia laevigata after a more comprehensive sampling . In addition to geographic clustering, Ophiactis savignyi, Ophionereis porrecta and Ophiolepis cincta are represented by specimens living in the same region but showing COI sequences that were assigned to two to three different putative species (cases of putative sympatry in the Red Sea and Hawaii for the first species, Lizard Island, Australia for the second, and Nosy-Be, Madagascar, for the third one) . Also, due to the evolutionary history of some species, divergent mitochondrial lineages can coexist within a species. This was shown for Ophiura sarsii Lütken, 1855 from the Barent Sea where distances up to 3.11% have been reported among COI sequences . Therefore, the putative species determined here based on COI need to be evaluated with integrative taxonomy, taking into account the life cycle, ecology, and if necessary, additional DNA data in order to infer the evolutionary history of the species [23, 85].
In a few cases, we were able to flag possible misidentifications in BOLD. These should be verified before considering the possibility of overlooked species diversity. Sequences of Dichrometra palmata (GBEH4375-13), Astroboa nuda (ECHOZ007-09) and Holothuria difficilis (GBMIN138025-18) were found in putative species that are not only separated from the conspecifics sequenced here and available in BOLD, but also grouped with species that are well differentiated based on both morphology and DNA data (Stephanometra indica for the first one, Holothuria pardalis and Astroglymma sculptum, Ophiocrene aenigma and Astroboa globifera for the second one). Our results also support the indication in BOLD that one specimen labeled as Heterocentrotus trigonarius (ECLI011-08) was possibly misidentified. Other misidentifications or errors are also possible, especially for singletons or when COI sequences of related species have not yet been sequenced (e.g. Holothuria atra BCUH041-09, GBEH8913-19, H. edulis GBEH8922-19 and Ophiothrix purpurea ECHOZ070-09).
Morphospecies not matching described species
Among the 18 morphospecies with a provisional name, 12 were grouped with species already described (either in the same putative species, or using the BOLD identification engine, cf. Table 4) and may thus represent yet uncharacterized phenotypes (either adults or juveniles) belonging to these described species. In the case of Comanthus sp., however, the distances to the closest species, Comanthus suavia and Comanthus parvicirrus (1.76–4.12%) overlap the range of distances between these two species (2.94–6.47%). Our specimen may represent a separate species of Comanthus because most current nominal species of this genus are widely distributed, include numerous synonyms and forms that are difficult to assign to phenotypic or interspecific variation . The COI sequences of seven morphospecies (including Aquilonastra cf. rowleyi that is split in two putative species) were separated from all BOLD sequences with a species identification. These morphospecies may correspond to species that have already been described but for which COI sequences are not available. Indeed, 16% of all known echinoderm species have not been barcoded yet according to the barcode of life initiative . Else, they may correspond to species that have not yet been described, as suggested by the close match with morphospecies sequences also labelled with a provisional name by independent teams (e.g. Aquilonastra sp.2 in Table 4). One juvenile specimen provisionally named as Amphiuridae sp. was grouped with a BOLD record identified as Ophiocoma cf. doederleini, which belongs to the Ophiocomidae. It is possible that juveniles of Amphiuridae and Ophiocomidae share some morphological characters that become family-specific only at the adult stage.
By using a simple DNA barcoding approach, our study contributed to improve our knowledge on the echinoderm biodiversity found along the East Coast of South Africa. This standard method supported by the BOLD system enables the analysis of many samples collected by different research teams at broad biogeographic scales. It confirmed, and sometimes revealed, species that should be investigated further in order to improve their delineation and describe a diversity that was not noticed so far. This is of primary importance because echinoderms represent a considerable biomass in marine habitats, play a major role in the marine ecosystems and have an economic value in the food and medical industries. The numerous putative species defined here based on COI sequences need to be better characterized and integrated in taxonomy. For this, sequencing mitochondrial and nuclear DNA from type material would provide valuable information to assign formal species names to these COI-based putative species. Keeping a sound connection between taxonomy and DNA data is of primary importance if future DNA-based investigations are expected to rely on the wide range of observations that are continuously reported for formally described species.
S1 Table. Specimen information of the 312 echinoderms specimens analysed in this study.
Columns ABGD, GMYC, bPTP and BIN provide the putative species IDs obtained with each method. Haplotype IDs are given for the representative sequences included in the phylogenetic tree reconstruction.
S1 File. Majority-rule consensus tree generated using a Bayesian inference and based on the COI dataset of Asteroidea.
The tree is in NEXUS format. Values at nodes represent posterior probabilities and tip labels include haplotype IDs, species name, origin of the data (previously in BOLD or generated in this study (“NEW”) and potential species IDs according to the species delimitation methods ABGD, GMYC, bPTP and BIN. Multiple values present at the same tip are provided and separated by a ‘-‘ (for example “BOLD-NEW” means that the haplotype is represented by sequences both recovered from BOLD and sequenced here).
S2 File. Majority-rule consensus tree generated using a Bayesian inference and based on the COI dataset of Crinoidea.
The tree is in NEXUS format. Nodes and tip labels are annotated as in S1 File.
S3 File. Majority-rule consensus tree generated using a Bayesian inference and based on the COI dataset of Echinoidea.
The tree is in NEXUS format. Nodes and tip labels are annotated as in S1 File.
S4 File. Majority-rule consensus tree generated using a Bayesian inference and based on the COI dataset of Holothuroidea.
The tree is in NEXUS format. Nodes and tip labels are annotated as in S1 File.
We thank Prof. Ahmed S. Thandar (School of Life Sciences, Biology, University of KwaZulu-Natal, South Africa) and Dr. Jennifer Olbers, (WILDTRUST, Durban, KwaZulu-Natal, South Africa) for their help in interpreting the results and double-checking morphological identifications. We also thank the contributions of Didier Van den Spiegel (RMCA) for the elaboration of the original project, Brigitte Segers (RBINS) for the preparation of the tissue samples and Dr. Massimiliano Virgilio (RMCA) and Zoë Decorte (RMCA) for the DNA analysis.
- 1. Pawson DL. Phylum Echinodermata. Zootaxa. 2007;1668:749–64.
- 2. Mah CL. New genera, species and occurrence records of Goniasteridae (Asteroidea; Echinodermata) from the Indian Ocean. Zootaxa. 2018 Dec 21;4539(1):1–116. pmid:30647336
- 3. Mooi R, Noordenburg H van. A new species and comparative morphology of Philippine sea biscuits (Echinoidea: Clypeaster). Zootaxa. 2021 Apr 21;4964(1):1–36.
- 4. Martins L, Tavares M. A new species of Parathyone (Holothuroidea: Dendrochirotida: Cucumariidae) from northeastern Brazil, with a key to species. Zootaxa. 2021 Jun 14;4985(2):245252.
- 5. Solís-Marín FA, Conejeros-Vargas CA, Caballero-Ochoa AA, Arriaga-Ochoa JA. Epitomapta simentalae sp. n., a new species of apodous sea cucumber from the Central Eastern Pacific coast of Mexico (Echinodermata, Holothuroidea, Apodida). Zookeys. 2019 Jan 15;817:1–9.
- 6. Messing CG. Three new species of Comasteridae (Echinodermata, Crinoidea) from the tropical western Pacific. Zoosystema. 2003 Dec 31;25(1):149–62.
- 7. Panina EG, Stepanov VG, Martynov A v. Psolidium kharlamenkoi—a new species of holothurian (Holothuroidea: Dendrochirotida: Psolidae) from the Kuril Islands. Zootaxa. 2021 Jun 30;4995(2):266–80.
- 8. Hebert PDN, Cywinska A, Ball SL, deWaard JR. Biological identifications through DNA barcodes. Proc Biol Sci. 2003;270(1512):313–21. pmid:12614582
- 9. Layton KKS, Corstorphine EA, Hebert PDN. Exploring canadian echinoderm diversity through DNA barcodes. PLoS One. 2016;11(11):1–16. pmid:27870868
- 10. Ward RD, Holmes BH, O’Hara TD. DNA barcoding discriminates echinoderm species. Mol Ecol Resour. 2008;8(6):1202–11. pmid:21586007
- 11. Laakmann S, Boos K, Knebelsberger T, Raupach MJ, Neumann H. Species identification of echinoderms from the North Sea by combining morphology and molecular data. Helgol Mar Res. 2016;70(1):18.
- 12. Bribiesca-Contreras G, Solís-Marín FA, Laguarda-Figueras A, Zaldívar-Riverón A. Identification of echinoderms (Echinodermata) from an anchialine cave in Cozumel Island, Mexico, using DNA barcodes. Mol Ecol Resour. 2013;13(6):1137–45. pmid:23551841
- 13. Ubagan MD, Lee T, Kim P, Shin S. A new species of the genus Henricia (Asteroidea, Spinulosida, Echinasteridae) from South Korea. Zookeys. 2020 Nov 25;997:1–15.
- 14. Eléaume M, Hemery LG, Bowden DA, Roux M. A large new species of the genus Ptilocrinus (Echinodermata, Crinoidea, Hyocrinidae) from Antarctic seamounts. Polar Biol. 2011 Sep 3;34(9):1385–97.
- 15. Boissin E, Hoareau TB, Paulay G, Bruggemann JH. Shallow-water reef ophiuroids (Echinodermata: Ophiuroidea) of Réunion (Mascarene Islands), with biogeographic considerations. Zootaxa. 2016 Apr 6;4098(2):273–97. pmid:27394586
- 16. Yamana Y, Nakaguchi K, Yamaguchi S, Katoh M, Ogawa A, Ohtsuka S. Four new dendrochirotid holothurians collected from the Seto Inland Sea and the western part of the Sea of Japan, western Japan. Zootaxa. 2021 Aug 17;5023(1):1–43. pmid:34810979
- 17. Rodriguez A, Hernandez JC, Clemente S, Coppard SE. A new species of Diadema (Echinodermata: Echinoidea: Diadematidae) from the eastern Atlantic Ocean and a neotype designation of Diadema antillarum (Philippi, 1845). Zootaxa. 2013 Apr 3;3636(1):144–70.
- 18. Williams ST. Species boundaries in the starfish genus Linckia. Mar Biol. 2000;136(1).
- 19. Crawford TJ, Crawford BJ. Linckia multifora (Echinodermata: Asteroidea) in Rarotonga, Cook Islands: Reproductive Mechanisms and Ecophenotypes. Pac Sci. 2007;61(3):371–81.
- 20. Jossart Q, Sands CJ, Sewell MA. Dwarf brooder versus giant broadcaster: combining genetic and reproductive data to unravel cryptic diversity in an Antarctic brittle star. Heredity (Edinb). 2019;123:622–33.
- 21. Galaska MP, Sands CJ, Santos SR, Mahon AR, Halanych KM. Crossing the Divide: Admixture Across the Antarctic Polar Front Revealed by the Brittle Star Astrotoma agassizii. Biol Bull. 2017 Jun;232(3):198–211.
- 22. Stöhr S, Weber AAT, Boissin E, Chenuil A. Resolving the Ophioderma longicauda (Echinodermata: Ophiuroidea) cryptic species complex: five sisters, three of them new. Eur J Taxon. 2020 Feb 4;(600):1–37.
- 23. Weber AAT, Stöhr S, Chenuil A. Species delimitation in the presence of strong incomplete lineage sorting and hybridization: Lessons from Ophioderma (Ophiuroidea: Echinodermata). Mol Phylogenet Evol. 2019 Feb 1;131:138–48.
- 24. Griffiths CL, Robinson TB, Lange L, Mead A. Marine biodiversity in south africa: An evaluation of current states of knowledge. PLoS One. 2010;5(8):e12008. pmid:20689849
- 25. Ratnasingham S, Hebert PDN. The Public Data Portal of the Barcode of Life Data System (www.barcodinglife.org), visited on 24 May 2022.
- 26. Roberts CM, McClean CJ, Veron JEN, Hawkins JP, Allen GR, McAllister DE, et al. Marine Biodiversity Hotspots and Conservation Priorities for Tropical Reefs. Science (1979). 2002 Feb 15;295(5558):1280–4. pmid:11847338
- 27. Clark AM, Rowe FWE. Monograph of shallow-water Indo-West Pacific echinoderms: i-vii, 1–238, pls 1–31. London: Trustees of the British Museum (Natural History); 1971. 238 p.
- 28. Clark AM, Courtman-Stock J. The Echinoderms of Southern Africa. London: British Museum (Natural History); 1976. 277 p.
- 29. Horton T, Kroh A, Ahyong S, Bailly N, Bieler R, Boyko CB, et al. World Register of Marine Species (WoRMS) [Internet]. WoRMS Editorial Board; 2022. Available from: https://www.marinespecies.org
- 30. QGIS Association. QGIS 3.22. Geographic Information System Developers Manual [Internet]. 2022 [cited 2022 Aug 8]. Available from: https://docs.qgis.org/3.22/en/docs/developers_guide/index.html
- 31. GEBCO Compilation Group. GEBCO 2020 Grid. 2020.
- 32. Natural Earth. Free vector and raster map data [Internet]. [cited 2022 Aug 8]. Available from: naturalearthdata.com
- 33. Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol Mar Biol Biotechnol. 1994;3(5):294–9. pmid:7881515
- 34. Arndt A, Marquez C, Lambert P, Smith MJ. Molecular Phylogeny of Eastern Pacific Sea Cucumbers (Echinodermata: Holothuroidea) Based on Mitochondrial DNA Sequence. Mol Phylogenet Evol. 1996 Dec 1;6(3):425–37. pmid:8975697
- 35. Lobo J, Costa PM, Teixeira MAL, Ferreira MSG, Costa MH, Costa FO. Enhanced primers for amplification of DNA barcodes from a broad range of marine metazoans. BMC Ecol. 2013;13:1–8.
- 36. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol [Internet]. 1990 Oct;215(3):403–10. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0022283605803602 pmid:2231712
- 37. Ratnasingham S, Hebert PDN. BOLD: The Barcode of Life Data System (www.barcodinglife.org). Mol Ecol Notes. 2007;7:355–64.
- 38. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004 Mar;32(5):1792–7. pmid:15034147
- 39. Kumar S, Stecher G, Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol Biol Evol. 2016;33(7):1870–4. pmid:27004904
- 40. Paradis E, Claude J, Strimmer K. APE: Analyses of Phylogenetics and Evolution in R language. Bioinformatics. 2004;20(2):289–90. pmid:14734327
- 41. Puillandre N, Lambert A, Brouillet S, Achaz G. ABGD, Automatic Barcode Gap Discovery for primary species delimitation. Mol Ecol. 2012 Apr;21(8):1864–77. pmid:21883587
- 42. Ratnasingham S, Hebert PDN. A DNA-Based Registry for All Animal Species: The Barcode Index Number (BIN) System. Fontaneto D, editor. PLoS One 2013 Jul 8;8(7):e66213. pmid:23861743
- 43. Pons J, Barraclough TG, Gomez-Zurita J, Cardoso A, Duran DP, Hazell S, et al. Sequence-Based Species Delimitation for the DNA Taxonomy of Undescribed Insects. Hedin M, editor. Syst Biol. 2006 Aug 1;55(4):595–609.
- 44. Zhang J, Kapli P, Pavlidis P, Stamatakis A. A general species delimitation method with applications to phylogenetic placements. Bioinformatics. 2013 Nov 15;29(22):2869–76. pmid:23990417
- 45. Kekkonen M, Hebert PDN. DNA barcode‐based delineation of putative species: efficient start for taxonomic workflows. Mol Ecol Resour. 2014 Jul 10;14(4):706–15. pmid:24479435
- 46. abgd web [Internet]. [cited 2021 Jan 13]. Available from: https://bioinfo.mnhn.fr/abi/public/abgd/abgdweb.html
- 47. Ezard T, Fujisawa T, Barraclough T. splits: SPecies’ LImits by Threshold Statistics. 2009.
- 48. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, et al. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012 May;61(3):539–42. pmid:22357727
- 49. Ferreira MAR, Suchard MA. Bayesian analysis of elapsed times in continuous-time Markov chains. Can J Stat. 2008;36(3):355–68.
- 50. Miller MA, Pfeiffer W, Schwartz T. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. In: Proceedings of the Gateway Computing Environments Workshop (GCE), 14 Nov 2010, New Orleans, LA. 2010. p. 1–8.
- 51. Lanfear R, Calcott B, Kainer D, Mayer C, Stamatakis A. Selecting optimal partitioning schemes for phylogenomic datasets. BMC Evol Biol. 2014;14(1):82. pmid:24742000
- 52. Paradis E. pegas: an R package for population genetics with an integrated-modular approach. Bioinformatics. 2010 Feb 1;26(3):419–20. pmid:20080509
- 53. Kruschke JK. Doing Bayesian data analysis: A tutorial with R, JAGS, and Stan, second edition. Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan, Second Edition. 2014.
- 54. Meier R, Blaimer BB, Buenaventura E, Hartop E, Rintelen T, Srivathsan A, et al. A re‐analysis of the data in Sharkey et al.’s (2021) minimalist revision reveals that BINs do not deserve names, but BOLD Systems needs a stronger commitment to open science. Cladistics. 2021 Sep 6;38:264–75. pmid:34487362
- 55. Jin Q, Hu XM, Han HL, Chen F, Cai WJ, Ruan QQ, et al. A two-step DNA barcoding approach for delimiting moth species: moths of Dongling Mountain (Beijing, China) as a case study. Sci Rep. 2018 Dec 24;8(1):14256. pmid:30250036
- 56. Dellicour S, Flot JF. The hitchhiker’s guide to single-locus species delimitation. Mol Ecol Res. 2018 Nov;18(6):1234–46. pmid:29847023
- 57. O’Hara TD, Hugall AF, Cisternas PA, Boissin E, Bribiesca-Contreras G, Sellanes J, et al. Phylogenomics, life history and morphological evolution of ophiocomid brittlestars. Mol Phylogenet Evol. 2019 Jan;130:67–80. pmid:30308280
- 58. Torrence KG, Correia MD, Hoffman EA. Divergent sympatric lineages of the Atlantic and Indian Ocean crinoid Tropiometra carinata. Invertebr Biol. 2012 Dec;131(4):355–65.
- 59. Pentinsaari M, Vos R, Mutanen M. Algorithmic single-locus species delimitation: effects of sampling effort, variation and nonmonophyly in four methods and 1870 species of beetles. Mol Ecol Resour. 2017 May;17(3):393–404. pmid:27292571
- 60. Lessios HA, Kessing BD, Pearse JS. Population structure and speciation in tropical seas: global phylogeography of the sea urchin Diadema. Evolution (N Y). 2001;55(5):955–75.
- 61. Geyer LB, Zigler KS, Tiozzo S, Lessios HA. Slow evolution under purifying selection in the gamete recognition protein bindin of the sea urchin Diadema. Sci Rep. 2020;10(1):9834.
- 62. Mita M, Uehara T, Nakamura M. Speciation in four closely related species of sea urchins (genus echinometra) with special reference to the acrosome reaction. Invertebr Reprod Dev. 2004;45(3):169–74.
- 63. Palumbi SR, Grabowsky G, Duda T, Geyer L, Tachino N. Speciation and population genetic structure in tropical Pacific Sea urchins. Evolution (N Y). 1997;51(5):1506–17. pmid:28568622
- 64. Arakaki Y, Uehara T. Morphological comparison of black Echinometra individuals among those in the Indo-West Pacific. Zoolog Sci. 1999;16(3):551–8.
- 65. Bronstein O, Loya Y. The Taxonomy and Phylogeny of Echinometra (Camarodonta: Echinometridae) from the Red Sea and Western Indian Ocean. PLoS One. 2013;8(10):e77374.
- 66. Ketchum RN, DeBiasse MB, Ryan JF, Burt JA, Reitzel AM. The complete mitochondrial genome of the sea urchin, Echinometra sp. EZ. Mitochondrial DNA B Resour. 2018;3(2):1225–7.
- 67. Jangoux M. Les Astéries de l’Ile Inhaca (Mozambique)(Echinodermata, Asteroidea). Annalen Koninklijk Museum voor Midden-Afrika—Zoologische wetenschappen. 1973;208:1–50.
- 68. Inoue J, Hisata K, Yasuda N, Satoh N. An Investigation into the Genetic History of Japanese Populations of Three Starfish, Acanthaster planci, Linckia laevigata, and Asterias amurensis, Based on Complete Mitochondrial DNA Sequences. G3 (Bethesda). 2020 Jul 1;10(7):2519–28.
- 69. Lessios HA, Kane J, Robertson DR. Phylogeography of the pantropical sea urchin Tripneustes: Contrasting patterns of population structure between oceans. Evolution (N Y). 2003;57(9):2026–36.
- 70. Ahmed Q, Thandar AS, Ali QM. Holothuria (Lessonothuria) insignis Ludwig, 1875 (formally resurrected from synonymy of H. pardalis Selenka, 1867) and Holothuria (Lessonothuria) lineata Ludwig, 1875—new additions to the sea cucumber fauna of Pakistan, with a key to the subgenus Lessonothuria Deichmann (Echinodermata: Holothuroidea). Zootaxa. 2020 Apr 24;4767(2).
- 71. Samyn Y, Massin C, Vandenspiegel D. The sea cucumber Holothuria lineata Ludwig, 1875 (Holothuroidea, Aspidochirotida, Holothuriidae) re-described from the newly found type. Zookeys. 2019 Apr 8;836:81–91. pmid:31048959
- 72. Borrero-Pérez GH, Vanegas-González MJ. Holothuria (Mertensiothuria) viridiaurantia sp. nov. (Holothuriida, Holothuriidae), a new sea cucumber from the Eastern Pacific Ocean revealed by morphology and DNA barcoding. Zookeys. 2019 Dec 2;893.
- 73. Michonneau F. Cryptic and not-so-cryptic species in the complex Holothuria (Thymiosycia) impatiens (Forsskål, 1775) (Echinodermata: Holothuroidea). bioRxiv [Preprint] [Internet]. 2015 [cited 2022 Jun 2];014225. Available from: https://www.biorxiv.org/content/10.1101/014225v1.full
- 74. Ahmed MI, Aamer Maher A, Lawrence AJ. Identification of the Holothurian species of the Red Sea and Gulf of Aqaba using DNA barcoding technique. Egypt J Aquat Biol Fish. 2016 Oct 1;20(4):1–7.
- 75. Michonneau F, Borrero-Perez G H, Honey M, Kamarudin KR, Kerr AM, Kim S, et al. The littoral sea cucumbers (Echinodermata: Holothuroidea) of Guam re-assessed–a diversity curve that still does not asymptote. Cah Biol Mar. 2013;54:531–40.
- 76. Samyn Y, Vandenspiegel D, Massin C. A new Indo-West Pacific species of Actinopyga (Holothuroidea: Aspidochirotida: Holothuriidae). Zootaxa. 2006 Mar 2;1138(1):53.
- 77. Boissin E, Hoareau TB, Paulay G, Bruggemann JH. DNA barcoding of reef brittle stars (Ophiuroidea, Echinodermata) from the southwestern Indian Ocean evolutionary hot spot of biodiversity. Ecol Evol. 2017 Dec;7(24):11197–203. pmid:29299292
- 78. Hoareau TB, Boissin E, Paulay G, Bruggemann JH. The Southwestern Indian Ocean as a potential marine evolutionary hotspot: perspectives from comparative phylogeography of reef brittle-stars. J Biogeogr. 2013;40(11):2167–79.
- 79. Taboada S, Pérez-Portela R. Contrasted phylogeographic patterns on mitochondrial DNA of shallow and deep brittle stars across the Atlantic-Mediterranean area. Sci Rep. 2016 Sep 2;6(1):32425. pmid:27585743
- 80. Stöhr S, O’Hara TD, Thuy B. Global Diversity of Brittle Stars (Echinodermata: Ophiuroidea). PLoS One. 2012 Mar 2;7(3):e31940. pmid:22396744
- 81. Pérez-Portela R, Almada V, Turon X. Cryptic speciation and genetic structure of widely distributed brittle stars (Ophiuroidea) in Europe. Zool Scr. 2013 Mar;42(2):151–69.
- 82. Khodami S, Martinez Arbizu P, Stöhr S, Laakmann S. Molecular Species Delimitation of Icelandic Brittle Stars (Ophiuroidea). Pol Polar Res. 2014 Jul 29;35(2):243–60.
- 83. Crandall ED, Treml EA, Liggins L, Gleeson L, Yasuda N, Barber PH, et al. Return of the ghosts of dispersal past: historical spread and contemporary gene flow in the blue sea star Linckia laevigata. Bull Mar Sci. 2014 Jan 1;90(1):399–425.
- 84. Genelt-Yanovskiy E, Li Y, Stratanenko E, Zhuravleva N, Strelkova N, Xu Q, et al. Phylogeography of the Brittle Star Ophiura sarsii Lütken, 1855 (Echinodermata: Ophiuroidea) from the Barents Sea and East Atlantic. Diversity (Basel). 2021 Jan 21;13(2):40.
- 85. Hugall AF, O’Hara TD, Hunjan S, Nilsen R, Moussalli A. An Exon-Capture System for the Entire Class Ophiuroidea. Mol Biol Evol. 2016 Jan;33(1):281–94. pmid:26474846
- 86. Summers MM, Messing CG, Rouse GW. The genera and species of Comatulidae (Comatulida: Crinoidea): taxonomic revisions and a molecular and morphological guide. Zootaxa. 2017 May 16;4268(2):151. pmid:28610370