Genetic Fingerprints in the Water: eDNA Outperforms Nets in Andean Fish Surveys

Deep in the mountains of northwestern Colombia, a team of researchers has shown that the water itself can tell us more about fish than fishing ever could. In a study published in Discover Conservation, scientists from the University of Antioquia and the University of the Americas tested environmental DNA metabarcoding against conventional capture methods in the reservoirs of the Porce River, a waterway transformed by cascading hydroelectric dams. Their conclusion is striking: by simply sampling water and reading the genetic traces shed by fish, they detected more species than nets, electrofishing, and hooks combined, all without killing or disturbing a single animal.

The stakes are high in this corner of South America. Colombia hosts 1,692 recorded freshwater fish species, 419 of which are found nowhere else on Earth. Many of these endemics live in waters fragmented by hydroelectric infrastructure, where dams block migration routes, alter flow and temperature, and isolate populations into shrinking fragments. Monitoring these communities has traditionally meant electrofishing, deploying gill nets, or using anesthetic and toxic substances, methods that are labor-intensive, expensive, and often lethal. For species already in decline, the very act of counting them can push them closer to the edge.

Environmental DNA, or eDNA, offers a radically different approach. Every fish constantly releases genetic material into its surroundings through skin cells, mucus, waste, and gametes. These DNA fragments persist in water for hours to days, and by filtering water samples and amplifying short, standardized genetic barcodes, researchers can identify which species were recently present without ever seeing them. Since the pioneering work demonstrating that eDNA could detect freshwater organisms, the technique has exploded in popularity, propelled by high-throughput sequencing and increasingly sophisticated bioinformatics. Several European countries, including Denmark, Finland, and Iceland, have already written eDNA monitoring into law under the European Water Framework Directive.

The Colombian team set out to bring this toolkit to the Andean Neotropics, a region where eDNA application has lagged despite its extraordinary biodiversity. Their field site was the Porce III reservoir, part of a chain of dams on the Porce River, which rises at 2,660 meters above sea level and runs 227 kilometers before joining the Nechi River. The Porce II plant, operating since 2002, impounds 142.7 million cubic meters of water across 890 hectares; Porce III, established downstream in 2011, floods another 461 hectares. Between 2019 and 2020, the researchers collected 25 water samples from eight monitoring sites spanning rivers, streams, and reservoir habitats, across both dry and rainy seasons, while simultaneously running traditional fish surveys at every site.

The molecular work involved a series of technical decisions that proved decisive. The team began with a general metazoan primer pair targeting the mitochondrial COX1 gene, but in silico simulations revealed that this primer set would require tolerating at least seven mismatches to amplify local fish DNA, most of them in non-fish vertebrate sequences. Field results confirmed the problem: only three fish species were detected, representing a mere 0.17 percent of total reads. The researchers responded by designing a modified COX1 primer, FishF2, tailored to native fish families of the Magdalena-Cauca basin, and by developing an entirely new primer pair, 300enVF/300enVR, targeting the mitochondrial 12S ribosomal gene. Computational screening of 2,624 mitochondrial genomes from 2,603 fish species identified the 12S pair as the clear winner, with 57.4 percent specificity for fish DNA and an 84 percent probability of recovering sequences identifiable to species level.

Calibration was equally rigorous. The team extracted genomic DNA from 175 tissue samples representing 38 known freshwater species held in the University of Antioquia’s ichthyology collection, pooled them, and created a ten-step serial dilution series. This mock community allowed them to tune the bioinformatic pipeline, built on the DADA2 denoising algorithm and the QIIME2 platform, to maximize true detections while suppressing false positives from sequencing errors and contamination. The final thresholds recovered 31 of the 38 known species. Crucially, field blanks and extraction blanks processed alongside every batch yielded no meaningful sequences, ruling out contamination as a confounding factor.

The head-to-head comparison delivered a clear verdict. Across the standardized campaigns, the 12S marker detected 23 fish taxa, traditional fishing methods recorded 14 species, and the optimized COX1 marker identified just 6 taxa. In one rainy-season survey, eDNA recovered nine taxa from a single water sample where fishing nets caught only two. In the final July 2020 campaign, COX1 failed to amplify entirely, yet 12S detected 16 fish taxa from 195,050 reads, all assigned to fish, while fishing recorded 11 species. When both markers were combined, eDNA metabarcoding proved not only more sensitive than capture methods but also more cost-effective, requiring less field time and fewer personnel than electrofishing or gill netting. Published comparisons suggest eDNA can cut monitoring costs by factors of two to ten, with the advantage growing as survey scale and species richness increase.

The study also revealed important nuances. Detection biases varied by taxonomic order: catfishes, the Siluriformes, were consistently recovered by all methods, whereas characins and knifefishes showed greater variability between techniques. Some species caught by nets never appeared in the eDNA record, likely because of low DNA shedding rates, rapid degradation in warm tropical water, primer-template mismatches, or gaps in reference databases. Conversely, several genera were recovered only through eDNA, suggesting either poor catchability with conventional gear or higher rates of DNA release and persistence. The lesson, the authors argue, is that neither method is complete on its own; the most comprehensive picture of fish diversity emerges when molecular and physical surveys are combined.

Perhaps the most consequential contribution is the protocol itself. The researchers have released a standardized, replicable workflow covering water collection, DNA extraction, primer selection, and bioinformatic analysis, complete with publicly available scripts, reference databases, and raw sequencing data on GitHub. They built local COX1 and 12S reference databases from voucher specimens, a critical step in a region where GenBank coverage of Neotropical fish remains patchy. For practitioners with limited resources, they recommend prioritizing the 12S marker for its specificity and resolution, provided curated reference databases exist, and they point to growing resources such as the Mare-MAGE database, the MitoFish pipeline, and a newly published Colombian freshwater fish reference dataset compiled from 1,270 voucher specimens.

What begins as a technical exercise in molecular ecology carries real weight for conservation policy. Reservoirs like Porce III are exactly the kind of impacted, continuously changing environments where long-term monitoring matters most, and where lethal sampling is hardest to justify. A cheap, non-invasive, repeatable method that can track fish communities across seasons and years opens the door to evaluating ecosystem recovery, detecting invasive species early, prioritizing protected areas, and monitoring responses to climate change in the species-rich waters of the Neotropics. The fish of the Andes have been leaving their signatures in the water all along; now, finally, science is learning to read them.

Subject of Research: Environmental DNA metabarcoding for fish biodiversity monitoring in Andean river reservoirs

Article Title: Andean fish inventories using an eDNA metabarcoding approach

Article References: Andean fish inventories using an eDNA metabarcoding approach. (n.d.). https://doi.org/10.1007/s44353-026-00081-3

Image Credits: AI Generated

DOI: 10.1007/s44353-026-00081-3

Keywords: environmental DNA, eDNA metabarcoding, Andean fish, Colombia, freshwater biodiversity, 12S ribosomal gene, COX1, conservation, hydroelectric reservoirs, Neotropics, molecular monitoring, ichthyology