Combining DNA barcoding and morphology to identify larval fishes from the nearshore environment off the south-east coast of South Africa

The early life history stages of most marine fish species are undescribed. The problem is, most of these fishes have pelagic larvae which are minute, delicate forms. Linking the larval stage to an adult counterpart is extremely challenging as larvae are morphologically different from the adults. Historically, larval fish identification relied solely on distinguishing morphological characteristics and meristic measurements, which has resulted in taxonomic confusion and misidentification. The introduction of the deoxyribonucleic acid (DNA) barcoding technique as an alternative approach has been successful in positively identifying larval fishes. The correct identification of larval specimens is the key to a better understanding of larval ecology, which underpins the success of any adult fish population. This study aimed to positively identify larval fishes of the south-east coast of South Africa using morphological characteristics and DNA barcoding. Larval and eggs specimens for this study were collected from the shallow nearshore waters of the south-east coast of South Africa. A total of 177 larval specimens were used for morphological analysis. Body shape, gut shape, pigmentation and morphometric measurements (such as body depth, preanal length and total body length) were used to identify each specimen to the family level. In addition, a fragment of mitochondrial cytochrome c oxidase subunit 1 gene (COI) was adopted for sequencing to identify larval fish specimens and fish eggs. Sequences generated from this study were compared to those in the Barcode of Life Database (BOLD). When there were no close matches to a sequence, the GenBank nucleic acid sequence database, maintained by the National Center for Biotechnology Information (NCBI), was used as an alternative. A total of 18 different families were identified through morphology. Seventy-seven of the 177 larval specimens were not subjected to morphological identification due to physical damage. The majority of larvae identified using morphological characteristics belonged to either the Sparidae, Tripterygiidae or Gobiesocidae fish families. Through DNA barcoding, 12 fish families, 16 genera and 18 different species were identified. Ten DNA barcodes (categorised as ‘no match’) from 10 different larval specimens were not identified through any of the online databases. Therefore, the 2% threshold value was used to identify members of the same species. The K2P genetic distance relationships were calculated among the no match sequences and downloaded probability matches from NCBI. This resulted in two unknown specimens assigned to the Blenniidae and Gobiidae. All other taxa were identified to species level, except specimens representing the Gobiidae and Tripterygiidae families. Based on the K2P genetic distances Gobiidae representatives were categorised as members of the Caffrogobius genus. Twenty-eight barcodes represented specimens from the Tripterygiidae. DNA barcode data from COI was analysed using the standard phylogenetic procedures in MEGA6 to examine relationships and differentiation among sequences. These could not be identified to the lowest taxonomic rank due to limited sequence data to compare them with. The sequence data from these specimens gave different results in the two online databases. BOLD results were to family level (Tripterygiidae) and NCBI to the species level (Clinidae: Pavoclinus profundus). Results in this study confirmed the efficiency of the DNA barcoding technique in species level identification of fish larvae. The evidence from genetic barcodes of the Tripterygiidae specimens, supported by morphological characteristics, suggests the need for thorough research to identify the individuals to the species level. The fact that this study identified taxonomically problematic Gobiidae and Tripterygiidae specimens suggests that studies similar to this may highlight additional diversity and help to resolve the taxonomy of other species in these families. However, the lack of reference sequence data from the adult specimens, and especially those with cryptic diversity, were both shortcomings for the positive identification of larvae. With that being said, it shows the necessity for more research to be conducted on barcoding of larvae in general as to accommodate all kinds of species from biodiversity to economic perspectives