BackgroundThe increasing availability of reference libraries of DNA barcodes (RLDB) offers the opportunity to the screen the level of consistency in DNA barcode data among libraries, in order to detect possible disagreements generated from taxonomic uncertainty or operational shortcomings. We propose a ranking system to attribute a confidence level to species identifications associated with DNA barcode records from a RLDB. Here we apply the proposed ranking system to a newly generated RLDB for marine fish of Portugal.Methodology/Principal FindingsSpecimens (n = 659) representing 102 marine fish species were collected along the continental shelf of Portugal, morphologically identified and archived in a museum collection. Samples were sequenced at the barcode region of the cytochrome oxidase subunit I gene (COI-5P). Resultant DNA barcodes had average intra-specific and inter-specific Kimura-2-parameter distances (0.32% and 8.84%, respectively) within the range usually observed for marine fishes. All specimens were ranked in five different levels (A–E), according to the reliability of the match between their species identification and the respective diagnostic DNA barcodes. Grades A to E were attributed upon submission of individual specimen sequences to BOLD-IDS and inspection of the clustering pattern in the NJ tree generated. Overall, our study resulted in 73.5% of unambiguous species IDs (grade A), 7.8% taxonomically congruent barcode clusters within our dataset, but awaiting external confirmation (grade B), and 18.7% of species identifications with lower levels of reliability (grades C/E).Conclusion/SignificanceWe highlight the importance of implementing a system to rank barcode records in RLDB, in order to flag taxa in need of taxonomic revision, or reduce ambiguities of discordant data. With increasing DNA barcode records publicly available, this cross-validation system would provide a metric of relative accuracy of barcodes, while enabling the continuous revision and annotation required in taxonomic work.

BackgroundBuilding reference libraries of DNA barcodes is relatively straightforward when specifically designed primers are available to amplify the COI-5P region from a relatively narrow taxonomic group (e.g. single class or single order). DNA barcoding marine communities have been comparatively harder to accomplish due to the broad taxonomic diversity and lack of consistently efficient primers. Although some of the so-called “universal” primers have been relatively successful, they still fail to amplify COI-5P of many marine animal groups, while displaying random success even among species within each group. Here we propose a new pair of primers designed to enhance amplification of the COI-5P region in a wide range of marine organisms.ResultsAmplification tests conducted on a wide range of marine animal taxa, rendered possible the first–time sequencing of DNA barcodes from eight separated phyla (Annelida, Arthropoda, Chordata, Cnidaria, Echinodermata, Mollusca, Nemertea and Platyhelminthes), comprising a total of 14 classes, 28 orders, 57 families, 68 genus and 76 species.ConclusionsThese primers demonstrated to be highly cost-effective, which is of key importance for DNA barcoding procedures, such as for building comprehensive DNA barcode libraries of marine communities, where the processing of a large numbers of specimens from a wide variety of marine taxa is compulsory.

Young juvenile Solea senegalensis were exposed to three sediments with distinct contamination profiles collected from a Portuguese estuary subjected to anthropogenic sources of contamination (the Sado estuary, western Portugal). Sediments were surveyed for metals (cadmium, chromium, copper, nickel, lead and zinc), a metalloid (arsenic) and organic contaminants (polycyclic aromatic hydrocarbons, polychlorinated biphenyls and a pesticide, dichloro-diphenyl-trichloroethane plus its metabolites), as well as total organic matter, redox potential and particle fine fraction. The fish were exposed to freshly collected sediments in a 28-day laboratorial assay and collected for histological analyses at days 0 (T(0)), 14 (T(14)) and 28 (T(28)). Individual weighted histopathological indices were obtained, based on presence/absence data of eight and nine liver and gill pathologies, respectively, and on their biological significance. Although livers sustained more severe lesions, the sediments essentially contaminated by organic substances caused more damage to both organs than the sediments contaminated by both metallic and organic contaminants, suggesting a possible synergistic effect. Correlation analyses showed that some alterations are linked, forming distinctive histopathological patterns that are in accordance with the severity of lesions and sediment characteristics. The presence of large eosinophilic bodies in liver and degeneration of mucous cells in gills (a first-time described alteration) were some of the most noticeable alterations observed and were related to sediment organic contaminants. Body size has been found to be negatively correlated with histopathological damage in livers following longer term exposures. It is concluded that histopathological indices provide reliable and discriminatory data even when biomonitoring as complex media as natural sediments. It is also concluded that the effects of contamination may result not only from toxicant concentrations but also from their interactions, relative potency and sediment characteristics that ultimately determine bioavailability.

Morphology-based profiling of benthic communities has been extensively applied to aquatic ecosystems’ health assessment. However, it remains a low-throughput, and sometimes ambiguous, procedure. Despite DNA metabarcoding has been applied to marine benthos, a comprehensive approach providing species-level identifications for estuarine macrobenthos is still lacking. Here we report a combination of experimental and field studies to assess the aptitude of COI metabarcoding to provide robust species-level identifications for high-throughput monitoring of estuarine macrobenthos. To investigate the ability of metabarcoding to detect all species present in bulk DNA extracts, we contrived three phylogenetically diverse communities, and applied four different primer pairs to generate PCR products within the COI barcode region. Between 78–83% of the species in the contrived communities were recovered through HTS. Subsequently, we compared morphology and metabarcoding-based approaches to determine the species composition from four distinct estuarine sites. Our results indicate that species richness would be considerably underestimated if only morphological methods were used: globally 27 species identified through morphology versus 61 detected by metabarcoding. Although further refinement is required to improve efficiency and output of this approach, here we show the great aptitude of COI metabarcoding to provide high quality and auditable species identifications in estuarine macrobenthos monitoring.