Custom DNA primers reveal hidden zooplankton diversity in the Indian Ocean

Marine zooplankton sit at the base of the ocean’s food web, linking phytoplankton productivity to fish, whales, and everything in between. Yet when scientists try to take a genetic census of these drifting communities using DNA metabarcoding, some of the most ecologically important groups—copepods, krill, arrow worms, and hydrozoans—mysteriously vanish from the results. A new study published in Metabarcoding and Metagenomics shows why that happens and demonstrates a practical fix: custom-designed, taxon-specific DNA primers that coax these elusive groups out of hiding. When the researchers applied their new primer cocktails to plankton net samples from the Southwest Indian Ocean, the representation of target groups jumped three- to five-fold compared with a previous survey, and the combined assay identified 220 zooplankton species from just four net tows.

DNA metabarcoding has become a cornerstone technology for monitoring marine pelagic ecosystems. Instead of painstakingly sorting and identifying thousands of individual organisms under a microscope, researchers extract DNA from a bulk sample, amplify a short genetic marker, and sequence the resulting mixture on a high-throughput platform. The mitochondrial cytochrome c oxidase subunit I gene, or COI, is the preferred marker for zooplankton barcoding because it offers good species-level resolution and is backed by extensive reference databases such as the MetaZooGene Barcode Atlas, NCBI GenBank, and the Barcode of Life Database (BOLD). Metabarcoding can detect cryptic species, damaged specimens, and life stages like eggs and larvae that lack diagnostic features, making it invaluable for tracking biodiversity, ecosystem change, and climate impacts on ocean productivity.

But there is a catch. The sequencing platforms most commonly used for metabarcoding, such as the Illumina MiSeq, cannot reliably read the full-length COI barcode of roughly 650 base pairs designed by Folmer and colleagues in 1994. Scientists therefore rely on mini-barcodes—shorter fragments of 100 to 350 base pairs that retain enough variability for species-level identification. The workhorse for zooplankton is the universal miniCOI barcode of about 313 base pairs introduced by Leray and colleagues in 2013. The problem is that

The mismatch problem is fundamentally a matter of molecular geometry. Primers are short synthetic DNA strands, typically around twenty nucleotides long, that must latch onto conserved regions flanking the barcode fragment before the polymerase enzyme can copy the target sequence. When a primer binding site has accumulated mutations in a particular lineage—changes that may be silent for the organism but fatal for amplification—the primer anneals poorly or not at all, and that lineage simply fails to appear in the sequencing output. The result is not that the organisms were absent from the sample, but that the molecular fishing hook never caught them. This is precisely the phenomenon the Southwest Indian Ocean study set out to correct for copepods, euphausiids, chaetognaths, and hydrozoans, four groups that visual inspection of net samples showed to be abundant but whose genetic signal was being systematically suppressed by the universal miniCOI primer.

The scale of the reference data underlying the primer design effort is worth appreciating. The researchers assembled sequence datasets spanning 954 copepod species across 110 families and 345 genera, 65 euphausiid species, 25 chaetognath species concentrated in the Sagittoidea, and 404 hydrozoan species across 83 families and 200 genera, drawing from GenBank and BOLD as of December 2023. Wherever possible, two individuals per species from different geographic areas were included so that within-species variation would be captured before any primer-binding region was chosen. This matters because a primer designed against a single representative of a species risks failing on geographically divergent populations of the same animal, quietly reintroducing the very bias the exercise was meant to eliminate.

Choosing the best mini-barcode region among the hundreds of possible short fragments is a nontrivial computational problem, and the study’s approach illustrates how it can be done rigorously. A sliding-window analysis moved windows of 100 to 300 base pairs along each alignment in ten-base-pair increments, evaluating each candidate fragment for its mean Kimura 2-parameter genetic distance, the scarcity of zero pairwise distances between different species, and the degree to which evolutionary relationships inferred from the fragment matched those inferred from full-length sequences. Forty-two candidate fragments per group were then subjected to maximum-likelihood tree building, and the resulting comparison trees were quantitatively compared against reference trees built from full-length alignments using measures of both topology and branch-length divergence. Low scores in both metrics indicated that a short fragment preserved the phylogenetic signal of the full barcode, making it a defensible surrogate when sequencing read lengths demand brevity.

Confirming that a candidate region actually separates species is a separate hurdle from finding a variable one. The barcode gap concept, formalized in the barcode literature nearly two decades ago, requires that the maximum genetic distance within a species be smaller than the minimum distance between species; when those distributions overlap, a sequence cannot be assigned unambiguously. The study computed this gap directly by subtracting maximum intraspecific distance from minimum interspecific distance, and supplemented it with the Jeffries–Matusita statistic, a measure borrowed from remote sensing that tests whether two distributions of distances are statistically separable. Values of this statistic approach a ceiling of 1.414, with higher values indicating cleaner separation between intra- and interspecific variation. Only after passing these tests were primers designed for the flanking regions of each selected fragment.

One practical complication deserves emphasis: a single primer pair cannot always serve an entire high-level taxon. Flanking regions that are convenient for one family may be hopelessly divergent in another, so the researchers built primer cocktails—sets of multiple primers mixed in one reaction—rather than forcing a one-size-fits-all solution. For chaetognaths, whose evolutionary position among animal phyla is distant from the arthropods and cnidarians that dominate zooplankton biomass, two reverse primers were required. This cocktail strategy echoes recommendations from a recent literature review covering studies across the Atlantic, Pacific, Indian, Arctic, and Southern oceans, which documented repeated misperformance of the standard miniCOI primer combinations and explicitly urged pairing universal primers with complementary taxon-specific ones.

A subtle but important technical risk in any mitochondrial primer design is the possibility of amplifying numts, fragments of mitochondrial DNA that have been translocated into the nuclear genome over evolutionary time. Numts can masquerade as authentic mitochondrial sequences in metabarcoding outputs, inflating apparent species counts or producing phantom haplotypes. The study guarded against this by running each candidate primer through NCBI’s Primer-BLAST tool, verifying that the primers exclusively matched mitochondrial genes of the intended taxa rather than nuclear pseudogenes. This verification step is often omitted in rapid primer development but is essential for the integrity of downstream biodiversity estimates.

The biological logic behind read-count interpretation also merits explanation. In metabarcoding outputs, the number of sequencing reads attributed to a taxon is an imperfect but informative proxy for its relative biomass in the original sample. The finding that copepods and euphausiids dominated read counts in the four net-tow samples is consistent with their known ecological standing: copepods are typically the most abundant metazoans in the marine pelagic realm, and euphausiids such as krill form swarms that constitute critical forage for fish, seabirds, and baleen whales. That the genetic census recovered this expected biomass hierarchy, once primer bias was corrected, is itself evidence that the method was behaving sensibly rather than merely amplifying whatever amplified most easily.

The broader significance of this work lies in its contribution to global ocean observation frameworks. Zooplankton are recognized as an essential ocean variable, meaning they are deemed critical enough for sustained, standardized monitoring to inform assessments of ecosystem health and climate-driven change. Metabarcoding is increasingly central to such monitoring because it scales to the sample volumes and survey frequencies that conventional morphology-based identification cannot support. But a monitoring program is only as trustworthy as its least-biased measurement, and primer bias acts as a systematic error that can shift apparent community composition year after year. By demonstrating that targeted primers can raise the representation of under-detected groups three- to five-fold, the study provides a template that other regions and other under-sampled taxa can follow.

It is also worth noting what the approach does not solve. Incomplete barcode reference libraries remain a parallel bottleneck, particularly in understudied regions like the Southwest Indian Ocean, where recent efforts have added multiple new copepod records to public databases but coverage is still far from complete. A perfectly designed primer is of limited use if the sequence it recovers cannot be matched to a named species. The two problems—primer bias and reference library gaps—are complementary, and progress on metabarcoding fidelity requires attention to both. The primer design pipeline described here, from sliding-window screening through barcode-gap validation and numt avoidance, offers a reproducible workflow that can be adapted as reference databases grow, ensuring that the molecular census of the plankton becomes steadily more complete and more faithful to the communities actually drifting in the sea.

Subject of Research: Genetics

Article Title: The design and testing of taxon-specific mini-barcode markers for metabarcoding of marine zooplankton

Article References: Harpal, S., Groeneveld, J., Willows-Munro, S., Huggett, J., & Govender, A. (2025). The design and testing of taxon-specific mini-barcode markers for metabarcoding of marine zooplankton. Metabarcoding and Metagenomics, 9, Article e161301. https://doi.org/10.3897/mbmg.9.161301

Image Credits: AI Generated

DOI: 10.3897/mbmg.9.161301

Keywords: DNA metabarcoding, marine zooplankton, COI barcode, primer bias, Southwest Indian Ocean, biodiversity monitoring, copepods, krill, design, testing, taxon-specific, mini-barcode