The technique in question, DNA metabarcoding, works by amplifying short stretches of DNA with universal primers before sequencing. It is fast, inexpensive, and has matched the results of traditional monitoring methods such as bottom trawling and baited underwater cameras in numerous studies. Yet it carries a well-known flaw called primer bias. Because universal primers rarely bind equally well to every species’ DNA, some fish are amplified enthusiastically while others are amplified weakly or not at all, skewing the relative read counts that researchers rely on to estimate abundance.

To get around this, a team led by researchers in the Netherlands compared metabarcoding against PCR-free metagenomics, also known as native sequencing, in which all the DNA in a sample is sequenced directly without any amplification step. Both approaches used Oxford Nanopore Technologies’ sequencing platform, and both were applied to water samples collected from a controlled aquarium at Rotterdam Zoo and from two field settings in the Biesbosch nature reserve, where juvenile European sturgeon had recently been reintroduced. The study also tested a novel filtration method: instead of the conventional 1.2 micrometre disk filters, the team used coarse cigarette filters held in 3D-printed holders, which have a pore size roughly ten times larger.

In the controlled environment, both methods detected all seven fish species known to be present in the sturgeon aquarium. But in the field settings, metagenomics outperformed metabarcoding, detecting ten species in the Biesbosch samples compared to just five with metabarcoding. The researchers attribute this to species masking, a phenomenon in which the overwhelming abundance of one species’ DNA, in this case the European sturgeon, suppresses amplification of rarer species during PCR. The sturgeon contributed over 97 per cent of fish eDNA in the Biesbosch samples, potentially crossing a threshold where lesser species simply disappear from the metabarcoding readout.

The differences in relative read counts between the two methods were striking. In the aquarium, Tortonese’s stingray contributed 11 per cent of fish reads in metabarcoding but only 0.026 per cent in metagenomics, a 400-fold discrepancy. Conversely, the blonde ray was better represented in metagenomics than in metabarcoding. These patterns confirm that primer bias was indeed present in the primer sets used, and that PCR amplification is more effective for some species than for others. The researchers caution that translating relative read counts to actual abundance or biomass should be done with great care in both approaches.

Metagenomics also revealed something metabarcoding cannot: epigenetic information. Because Oxford Nanopore sequencing reads the native DNA molecule directly, it can detect chemical modifications such as methylation without additional experiments. The team identified three nucleotide positions on the 18S ribosomal RNA gene of the European sturgeon that showed different methylation patterns between eDNA samples. While the 18S gene is not directly linked to life-history traits, the finding serves as a proof of principle that a single metagenomic sequencing run can provide both species diversity and epigenetic patterns, information that would otherwise require separate bisulfite sequencing experiments.

The coarse cigarette filters, a surprisingly low-cost solution, showed mixed results. In the controlled environment, they recovered more species than conventional filters across all sequencing datasets, and in metagenomics they consistently outperformed the standard filters in both settings. However, in metabarcoding of field samples, the coarse filters did not improve species detection, possibly because most were not used to their maximum filtration capacity. A single coarse filter pushed to its limit, processing 8.3 litres of water, yielded a higher proportion of fish DNA than any other coarse filter in the study, hinting that maximum-volume filtration may be key to unlocking their potential.

The study also uncovered an unexpected complication: nuclear mitochondrial DNA fragments, or NUMTs, in the European sturgeon genome. These are copies of mitochondrial genes that have been inserted into the nuclear genome over evolutionary time. When amplified with mitochondrial primers, NUMTs produce misleading amplicons that can be mistaken for genuine mitochondrial DNA. In this study, NUMT-derived amplicons accounted for up to 40 per cent of COI metabarcoding data, potentially skewing relative read counts and complicating ecological interpretation. The researchers mapped these alternative amplicons to the sturgeon’s nuclear genome and confirmed their NUMT origin.

Cost remains a significant barrier for metagenomics. Sequencing a full PromethION flow cell cost roughly €60 per sample, five times the metabarcoding cost of about €12. But the researchers argue that metagenomic data contains information about the entire tree of life, not just fish, so for studies targeting multiple taxonomic groups, metagenomics may ultimately be more cost-efficient than running separate metabarcoding PCRs for each group.

The study’s authors conclude that metabarcoding and metagenomics function complementarily in biodiversity estimates. Metabarcoding excels at detecting rare species through the sheer sensitivity of PCR, while metagenomics circumvents primer bias and species masking, provides more representative relative abundance estimates, and adds epigenetic insights from the same sequencing run. Meanwhile, the coarse, low-cost cigarette filter and 3D-printed filter holder offer an affordable and accessible alternative for eDNA filtration that often outperforms conventional filters in revealing fish communities. As genomic reference databases continue to expand, particularly in understudied regions, the researchers expect metagenomics to play an increasingly central role in aquatic biodiversity monitoring.

Subject of Research: Comparative analysis of eDNA metabarcoding and eDNA metagenomics for fish biodiversity estimates using standard and novel filtration methods

Article Title: Comparative analysis of eDNA metabarcoding and eDNA metagenomics for fish biodiversity estimates using standard and novel filtration methods

Article References: van Berkel, D., Brevé, N. W. P., de Boer, M., Reynaud, E. G., & Nijland, R. (2026). Comparative analysis of eDNA metabarcoding and eDNA metagenomics for fish biodiversity estimates using standard and novel filtration methods. Metabarcoding and Metagenomics, 10, Article e182039. https://doi.org/10.3897/mbmg.10.182039

Image Credits: AI Generated

DOI: 10.3897/mbmg.10.182039

Keywords: Comparative, analysis, eDNA, metabarcoding, metagenomics, fish, biodiversity, estimates, standard, novel, filtration, methods