Biodiversity Data Journal :
Research Article
|
Corresponding author: Dirk Steinke (dsteinke@uoguelph.ca)
Academic editor: Pavel Stoev
Received: 24 Feb 2017 | Accepted: 12 Apr 2017 | Published: 13 Apr 2017
© 2017 Dirk Steinke, Jeremy deWaard, Martin Gomon, Jeffrey Johnson, Helen Larson, Oliver Lucanus, Glenn Moore, Sally Reader, Robert Ward
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Steinke D, deWaard J, Gomon M, Johnson J, Larson H, Lucanus O, Moore G, Reader S, Ward R (2017) DNA barcoding the fishes of Lizard Island (Great Barrier Reef). Biodiversity Data Journal 5: e12409. https://doi.org/10.3897/BDJ.5.e12409
|
To date the global initiative to barcode all fishes, FISH-BOL, has delivered barcodes for approximately 14,400 of the 30,000 fish species; there is still much to do to attain its ultimate goal of barcoding all the world’s fishes. One strategy to overcome local gaps is to initiate short but intensive efforts to collect and barcode as many species as possible from a small region – a barcode ‘blitz’. This study highlights one such event, for the marine waters around Lizard island in the Great Barrier Reef (Queensland, Australia). Barcode records were obtained from 983 fishes collected over a two-week period. The resulting dataset comprised 358 named species and another 13 species that presently can only be reliably identified to genus level. Overall, this short expedition provided DNA barcodes for 13% of all marine fish species known to occur in Queensland.
DNA barcoding, taxonomy, coral reef, biodiversity
DNA barcoding is a relatively new and powerful species identification tool that has found ready application to many animal taxa. It is based on sequencing a 650 bp region of the mitochondrial cytochrome c oxidase gene I (COI), the premise being that different species have different COI sequences or barcodes (
There have been several large-scale projects aimed at verifying its potential for discriminating fish (
One strategy that can be used to increase species coverage is a barcoding ‘blitz’ of a chosen region (
The Australian fish fauna is remarkably rich with many endemic species. Some 4600 species have been estimated for Australian waters, about 2300 of which have been recorded from the Great Barrier Reef (
In September 2008, a team of 12 Australian and Canadian scientists (see Acknowledgments) spent two weeks on the island, collecting, photographing and tissue sampling as many fish species as could be caught, under permit conditions provided by the Great Barrier Reef Marine Park Authority (GBRMPA). DNA sequencing was subsequently done at the Centre for Biodiversity Genomics in Guelph, Canada. We describe here the results of this barcoding blitz on the fishes of Lizard Island.
Members of the team arrived on Lizard Island on 3 September 2008 and departed 17-19 September. Fish were collected at different sites around Lizard Island under GBRMPA permit G26633, using clove oil, handnets, spears, light traps, beach seines, and hook and line approaches. Reefs, bommies, beaches, and mangroves were visited (Fig.
Prior to processing the specimens were morphologically identified by the appropriate expert using available taxonomic keys, field guides, and distribution records. Each specimen was assigned to one of five levels of reliability depending on the taxonomic expertise of the identifier involved and their intentions, following guidelines developed by Commonwealth Scientific and Industrial Research Organisation (CSIRO) fish taxonomists, and laid out in the FISH-BOL sampling protocol (
Level 1: highly reliable identification - specimen identified by an recognized authority of the group, or a specialist that is presently studying or has reviewed the group in the region in question.
Level 2: identification made with high degree of confidence at all levels - specimen identified by a trained identifier who had prior knowledge of the group in the region or used available literature to identify the specimen.
Level 3: identification made with high confidence to genus but less so to species - specimen identified by a trained identifier who was confident of its generic placement but did not substantiate their species identification using the literature, or a trained identifier who used the literature, but still could not make a positive identification to species, or an untrained identifier who used most of the available literature to make the identification.
Level 4: identification made with limited confidence - specimen identified by a trained identifier who was confident of its family placement, but unsure of generic or species identifications (no literature used apart from illustrations), or an untrained identifier who had/used limited literature to make the identification.
Level 5: identification superficial - specimen identified by a trained identifier who is uncertain ofthe family placement of the species (cataloging identification only), an untrained identifier using, at best, figures in a guide, or where the status and expertise of the identifier is unknown.
For this study, we collected 1,075 individuals, which were found to represent 395 fish species, from the waters around Lizard Island. When possible, several adults were analyzed per species. Specimens were kept on ice and subsequently imaged in the field (by convention left side of the animal). Samples used for DNA analysis were removed of lateral muscle from the right side of the specimen or by removing the right eye from very small specimens such as juveniles. Specimens are stored as vouchers in the Australian Museum, Sydney, the Australian National Fish Collection at CSIRO, Hobart, and the Western Australian Museum, Perth. Collection details are recorded in the public dataset DS-LIFE (http://dx.doi.org/10.5883/DS-LIFE) on the Barcode of Life Data Systems database (BOLD, http://www.boldsystems.org, see
DNA was extracted from the muscle tissue of each specimen using an automated glass Fiber (AcroPrep) method (
We used the analysis tools in BOLD to calculate the nucleotide composition of the sequences and distributions of Kimura-2-Parameter distances (
A list of all marine fish species (N=2764) currently present in Queensland was obtained from the Australian Faunal Directory in December 2016 (
Sequence data are available on both BOLD and GenBank. Specimen and collection data, sequences, specimen images, GenBank accession numbers, and trace files can be found in the public dataset DS-LIFE on BOLD (http://dx.doi.org/10.5883/DS-LIFE). An abbreviated version of the data is available in Suppl. material
For this study, we obtained 983 sequence records that derived from 358 named species (177 genera, 59 families) and another 13 species that could only be reliably identified to genus level (Table
Species name |
Order/Family |
BIN |
N |
Cheilodipterus cf. quinquelineatus |
Kurtiformes/Apogonidae |
1 |
|
Eviota sp. |
Gobiiformes/Gobiidae |
1 |
|
Eviota sp. 1 |
Gobiiformes/Gobiidae |
1 |
|
Eviota sp. 2 |
Gobiiformes/Gobiidae |
4 |
|
Eviota sp. 3 |
Gobiiformes/Gobiidae |
1 |
|
Eviota sp. 5 |
Gobiiformes/Gobiidae |
1 |
|
Gobiodon sp. |
Gobiiformes/Gobiidae |
1 |
|
Paragobiodon sp. |
Gobiiformes/Gobiidae |
1 |
|
Salarias sp. |
Blenniiformes/Blenniidae |
1 |
|
Scarus sp. |
Labriformes/Scaridae |
1 |
|
Scorpaena sp. |
Scorpaeniformes/Scropeanidae |
1 |
|
Trimma oki group 8 |
Gobiiformes/Gobiidae |
1 |
|
Trimmatom sp. |
Gobiiformes/Gobiidae |
2 |
All sequences met the quality (<1% N) and length (>500 bp) criteria for BIN assignment, and were assigned to 375 BINs. There was perfect correspondence between the specimens assigned to a particular BIN and the members of a particular morphospecies in nearly all cases (372 of 375). The three exceptions each involved a BIN split with the members of a particular species assigned to two BINs (Table
Discordances between BIN and species assignments (Species assigned to two BINs)
Species name |
Order/Family |
BIN 1 |
BIN 2 |
Amniataba caudavittata |
Perciformes/Terapontidae |
||
Ellochelon vaigiensis |
Mugiliformes/Mugilidae |
||
Gobiodon quinquestrigatus |
Gobiiformes/Gobiidae |
828 (84%) of the 983 barcoded specimens were correctly identified in the field by one of the fish taxonomists in the team (MG, JJ, PL, GM, SR, AH, see acknowledgements), 106 (11%) were initially misidentified, and 59 (6%) could not be identified to species level, receiving either a genus or family designation. Misidentifications were later exposed and resolved after DNA barcoding analysis and morphological re-inspection by experts with particular taxonomic knowledge. Identification errors in the field occurred more frequently when the identifier indicated a lower level of confidence, reflecting varying degrees of expertise (Fig.
The species detected in the present study were compared to the list of all fishes (N =2764) known from the marine waters of Queensland (
This study assembled DNA barcode sequences for 371 species of marine fishes that occur in the waters of Lizard Island (Table S1). This represents about 25% of the known marine ichthyofauna of the region (
Our study also revealed three cases of BIN splits involving the following taxa:
1. Amniataba caudavittata (
The five specimens of Yellowtail trumpeter A. caudavittata fell into two BINs (n=2, n=3) that show 5% sequence divergence between BINs but with novariation within BINs. This genus contains only three described species and one of those (A. perco) is barcode divergent (c. 7%) from both BINs. The other species (A. affinis) is known only from river systems and lagoons of Papua New Guinea and has not been barcoded. It seems likely that, despite no obvious morphological diversity, the two BINs comprise the original A. caudavittata and an overlooked cryptic species that requires description.
2. Ellochelon vaigiensis (
Specimens of E. vaigiensis were represented by two quite divergent (4.9%) BINs (n=5, n=1), which might reflect an instance of unjustified synonymization as several species (Mugil macrolepidotus, M. melanochir, M. tegobuan, M. occidentalis, M. ventricosus) have been recently synonymized under this species name (
3. Gobiodon quinquestrigatus (
Although COI divergence was quite low (1.24%), G. quinquestrigatus sequences were placed in two BINs (n=5, n=1). We were not able to find any morphological differences between members of these BINs nor any prior history of names-in-waiting. Without further evidence (e.g.,additional nuclear markers), this instance might either represent the discovery of a cryptic species or an artifact of the BIN algorithm due to high intraspecifc sequence variability and low sampling intensity. Valid species can harbour multiple mtDNA lineages with no morphological differences. With increased sampling such lineages can dissolve.
The speed of conducting this inventory reflected the team’s focus on a single group of organisms and the variety of collecting protocols deployed. One disadvantage of this type of fieldwork is the lack of time and resources available for proper initial identification of the difficult-to-delineate taxa. We identified samples in the field to one of five levels of reliability, depending on the taxonomic expertise of the identifier involved, following a standard protocol of the CSIRO Australia (see Materials and Methods, and
Although the project started as a rapid and intensive effort to collect and barcode as many species as possible from Lizard Island, it took several years to validate and confirm the species IDs. Some inventoried taxa still lack species-level determination, but these will be resolved over time. The barcodes obtained during this study, in concert with the BIN system of BOLD, facilitate crowd sourcing of the necessary taxonomic refinement (e.g.
Lizard Island is a unique natural reserve with the infrastructure necessary to conduct research on the northern section of the Great Barrier Reef, a barcode reference library and updated species inventory for its fishes adds to the infrastructure that can be shared with present and future researchers. This database is likely to become of increasing significance. In April 2014, Cyclone Ita passed directly across theisland – the most severe storm ever recorded for this location. The storms caused massive coral loss, further amplified by higher than average water temperatures in 2015 and 2016, which led to massive coral bleaching. The latter affected mostly the northern Great Barrier Reef, and one of the worst hit areas was around Lizard Island where about 90% of the coral died (
The Lizard Island team comprised, from Australia, Martin Gomon (Museum Victoria, Melbourne), Amanda Hay and Sally Reader (Australian Museum, Sydney), Jeff Johnson (Queensland Museum, Brisbane), Peter Last and Bob Ward (CSIRO National Research Collections Australia, Hobart) and Glenn Moore (Western Australian Museum, Perth), and from Canada, Jay Cossey, Jeremy deWaard, and Dirk Steinke (University of Guelph, Guelph), David Hardie (Dalhousie University, Halifax), and Oliver Lucanus (belowwater.com, Montreal). We warmly thank Ann Hoggett and Lyle Vail (Lizard Island Research Station) and Debra Moore (field volunteer). The expedition was made possible by the award of a grant from the Total Foundation (http://fondation.total) to Dirk Steinke for his application “DNA Barcoding of a Coral Reef Fish Community”. Species identifications were made in the field by members of the team. Subsequent corrections and final determinations were done with the additional help of Hiro Glenn, Ben Victor, William White, John Pogonoski, and John Randall. DNA barcode analysis was supported through funding provided by the Alfred P. Sloan Foundation (MarBOL), NSERC, the Canada Research Chairs program, and the government of Canada through Genome Canada.