Over the last decade, environmental DNA (eDNA) has emerged as a non-invasive and effective tool for characterizing species from water, soil, or air samples. The increasing interest in offshore and deep-sea explorations in Colombian oceans has required the implementation of new technology to improve sampling efforts and data quality in these hard-to-access areas.
In this study, we present the promising complementarity of eDNA-metabarcoding in providing accurate taxonomic delimitations fish obtained on board several scientific expeditions along the Colombian Pacific. We filtered 360L of seawater in each of the 34 sampling stations through 0.2µm capsule filters using peristaltic pumps. Seawater was collected during day and night shifts using oceanographic rosettes in vertical transects (200 to 10m depth). eDNA was preserved and analyzed using a 12S-V5 mitochondrial region, followed by high‑throughput sequencing. Molecular operational taxonomic units (MOTUs) were assigned using OBITools, as well as curated reference libraries that included regional voucher sequences.
The results yielded over 20 million DNA sequences, translated into 320 fish ASVs from 71 families, 59 genera and 84 species, most of which were migratory, demersal, meso-, and bathypelagic taxa. Families with the greatest species richness belonged to Teleostei families such as Myctophidae, Nomeidae, Scombridae, Coryphaenidae and Istiophoridae. Regarding Elasmobranchii, the most conspicuous families were Carcharhinidae, Alopiidae, Myliobatidae and Rhincodontidae. It is important to mention that another 100 MOTUS could not be identified to the species level due to lack of genetic reference databases available in the region.
These results were compared to historical data regarding fishing activities in the past 20 years, demonstrating that eDNA could match over 90% of the known fish biodiversity surveys in offshore areas while reducing sampling effort and costs. Further research should focus on improving genetic databases and on the implementation of eDNA tools as an alternative that should be embedded in routine monitoring programs.
