What Is DNA Barcoding and How Does It Work?

DNA barcoding is a method of identifying species by reading a short, standardized stretch of their genetic code, much the way a supermarket scanner reads the black-and-white stripes on a product label. A small tissue sample, a feather, a fin clip, or even a trace of shed cells is enough to extract DNA, amplify a specific gene region, and compare the resulting sequence against a reference library of known species.¹ The idea is deceptively simple, but the range of problems it solves stretches from catching seafood fraud to tracking endangered turtles through customs.

The Core Idea in Plain Terms

Every species accumulates small differences in its DNA over evolutionary time. DNA barcoding exploits the fact that these differences tend to be larger between species than within a single species. If you pick the right gene and sequence it from an unknown organism, the sequence will cluster tightly with others of its own kind and sit apart from sequences of different species. That gap between “how different individuals of one species look” and “how different two separate species look” is what makes the whole system work.

You do not need to sequence an organism’s entire genome. A fragment of just a few hundred base pairs from the right gene is enough to make a reliable identification in most cases. That short fragment is the “barcode.” The process is designed to be rapid, repeatable, and ultimately automatable, so that someone who is not a specialist taxonomist can still put a name on a specimen.

Which Gene Gets Read

There is no single barcode gene that works for all life. Different branches of the tree of life carry different genetic quirks, so researchers have settled on a small menu of markers tailored to each major group.

• Animals: The go-to gene is cytochrome c oxidase subunit I, usually shortened to COI (sometimes CO1 or cox1). It sits in the mitochondrial genome, evolves fast enough to distinguish closely related species, yet stays conserved enough that universal primers can latch onto it across a huge range of animal groups.² No other gene region currently has as many taxonomically verified sequences in public databases, which is a practical advantage when you need something to compare your unknown against.³
• Plants: Plant mitochondrial DNA evolves too slowly to separate species well, so botanists rely on chloroplast genes instead, chiefly rbcL and matK. These two are often used together, though combining them does not always improve identification compared with using one alone.⁴
• Fungi: The internal transcribed spacer region of ribosomal DNA, known as ITS, was formally proposed as the primary fungal barcode because it shows the clearest separation between species-level and within-species variation across the broadest range of fungal groups.⁵
• Bacteria and archaea: The 16S ribosomal RNA gene serves as the standard marker for these microorganisms.⁶

The Consortium for the Barcode of Life (CBOL) coordinated much of the early work in assigning these standard markers. The practical upshot is that when someone says “we barcoded it,” the gene they read depends on what kingdom the organism belongs to.

From Tissue Sample to Species Name

The laboratory workflow has a few well-defined steps, even though the details vary depending on the equipment at hand and the condition of the sample.

First, DNA is extracted from the specimen. Fresh tissue, blood, or a small piece preserved in ethanol works well. Standard commercial kits break open cells, remove proteins and debris, and leave behind purified DNA in solution.⁷ This step is not dramatically different from the DNA extraction you might have done with dish soap and rubbing alcohol in a high-school biology class, though the lab version is more controlled.

Next comes amplification. Using the polymerase chain reaction (PCR), researchers copy the target gene region millions of times over. This is necessary because the amount of DNA in a tiny tissue sample is too small to read directly. PCR uses short synthetic DNA fragments called primers that are designed to bind to the flanking regions of the barcode gene, ensuring that only the desired stretch gets copied.

The amplified DNA is then sequenced. For decades, the standard workhorse was Sanger sequencing, which reads one specimen at a time and produces a single, high-quality sequence. Sanger remains common, but next-generation sequencing platforms can process hundreds or thousands of specimens simultaneously, which speeds things up enormously and is starting to match or beat Sanger on cost.⁸ Long-read platforms like PacBio have shown higher success rates than Sanger for fungal barcoding at comparable cost, further expanding the toolkit.⁹

Finally, the sequence is compared against a reference database. If a close match exists, you get a species name. If nothing close comes back, you may have found something not yet represented in the library, which can itself be a valuable discovery.

The Reference Library That Makes It All Work

A barcode sequence is only useful if you have something to compare it against. The central repository for this purpose is the Barcode of Life Data System, or BOLD. It stores DNA barcode sequences alongside specimen photographs, collection data, and taxonomic information, creating a curated library that researchers worldwide can query.¹⁰ GenBank, the broader public sequence database run by NCBI, also holds barcode-region sequences, but BOLD was purpose-built for barcoding and enforces stricter quality controls on the voucher specimens behind each record.

BOLD’s complete COI library has grown to roughly 15.7 million records, which is both a strength and a logistical headache. Searching that many records for every query takes time, especially in large-scale environmental studies. A recently developed compressed version of the library, called BOLDistilled, pares the collection down to about 1.7 million representative records and cuts analysis time by over 98 percent without sacrificing the accuracy of species assignments.¹¹ That kind of optimization matters when researchers are processing thousands of environmental samples at a time.

Catching Food Fraud

One of the most publicly visible uses of DNA barcoding is exposing mislabeled food, especially seafood. Because a fish fillet does not look much like the species it came from, substitution is easy and profitable. Barcoding cuts through that ambiguity: sequence a sliver of the fillet, and the DNA reveals whether the “red snapper” on the menu really is red snapper.

A study that tracked seafood mislabeling in Los Angeles sushi restaurants from 2012 to 2015 found that nearly half of all samples tested were mislabeled. Halibut, red snapper, yellowfin tuna, and yellowtail were mislabeled more than 77 percent of the time, while salmon and mackerel were mislabeled much less often. Every sushi restaurant sampled had at least one case of mislabeling. High-end grocery stores selling sushi-grade fish fared only slightly better, at about 42 percent mislabeling.¹² A separate survey across 15 regions of Spain found half of food service establishments selling mislabeled seafood, with certain species like dusky grouper and tope shark mislabeled at rates between 86 and 100 percent.¹³

Beyond seafood, DNA barcoding is used to verify the identity of meat products and medicinal plants, where adulteration can carry health risks as well as economic consequences.¹⁴

Wildlife Forensics and Stopping Illegal Trade

When border agents intercept a suspicious shipment of animal products, the contents often look nothing like the species they came from. Ground meat, dried parts, processed traditional medicines: morphological identification is impossible. DNA barcoding gives enforcement agencies a tool that works on heavily processed material as long as enough DNA survives.

In one case from Pakistan, a consignment labeled as “fish meat” was intercepted at a port. When the samples were barcoded with fish-specific primers, the sequences matched at 99 percent to the Indian flap-shelled turtle, a species protected under the Convention on International Trade in Endangered Species (CITES).¹⁵ In South Africa, forensic casework has used COI and cytochrome b sequencing to identify confiscated material from illegally hunted antelope, matching samples to impala, eland, and kudu with sequence similarities above 99 percent.¹⁶ Cytochrome b fragments alongside COI are commonly used for mammalian species identification in forensic contexts, since both markers are well-represented in reference databases.¹⁷

Invasive Species and Biosecurity

Ports and airports are the front line for catching organisms that could devastate local ecosystems or agriculture if they get through. Traditional identification of intercepted insects relies on taxonomic specialists who may not be available around the clock or for every taxonomic group. DNA barcoding offers a more standardized alternative.

A reanalysis of tussock moth and fruit fly specimens intercepted at New Zealand’s border over a decade showed 90 and 96 percent agreement, respectively, between barcode-based identifications and the results of an older molecular method. Where the older method had left some tussock moth specimens as “unknowns,” barcoding placed them to family, genus, or species, and it resolved some fruit fly species complexes that had been lumped together.¹⁸

Once an invasive species has already established itself, early detection becomes the priority. Environmental DNA methods, where barcoding is applied to DNA traces shed into water or soil rather than to captured organisms, can detect invasive species at very low population densities and at any life stage. This has proven useful for tracking species like the American bullfrog in areas where conventional surveys would miss small or secretive populations.¹⁹

Environmental DNA and Metabarcoding

Standard DNA barcoding works on one specimen at a time: you take a piece of tissue, sequence it, and match it. Environmental DNA (eDNA) metabarcoding scales this up dramatically. Instead of catching or collecting organisms, researchers scoop up water, filter it, and extract whatever DNA happens to be floating in it — shed skin cells, mucus, feces, pollen, spores. Using universal primers and high-throughput sequencing, they amplify and read barcode regions from every species whose DNA is present in the sample simultaneously.²⁰

The breadth of detection is remarkable. A study applying multiple metabarcoding assays to seawater from a tropical marine environment picked up all major branches of life, including animals, fungi, protists, plants, chromists, bacteria, and archaea, from a single set of water samples.²¹ This makes eDNA metabarcoding a powerful tool for whole-ecosystem biodiversity snapshots without having to physically capture or even observe anything.

What Animals Eat, Traced Through Their Droppings

Dietary metabarcoding applies the same principle to animal feces instead of environmental water. By extracting and sequencing DNA from droppings, researchers can identify what an animal has been eating with much finer resolution than stomach-content dissection or field observation ever allowed.²² The method has been applied to species ranging from crustaceans and insects to whales.

An inventive study of insectivorous birds in northern Europe combined two angles: researchers barcoded nestling feces to identify the prey species the chicks had eaten, and barcoded caterpillar droppings (frass) collected from birch canopies to identify which prey species were actually available in the habitat. They found 53 prey species in the nestling feces, about a fifth of which also turned up in the frass samples, giving a direct view of how selective the birds were in their foraging.²³ This kind of paired analysis would be nearly impossible with traditional observation-based methods.

Portable Sequencing in the Field

Until recently, DNA barcoding required shipping samples back to a well-equipped laboratory, which could add days or weeks of delay. That is starting to change. Portable sequencers, especially the MinION device from Oxford Nanopore Technologies, are small enough to carry in a backpack and can produce sequence data in hours rather than days.

Researchers have demonstrated the full barcoding pipeline — DNA extraction, PCR amplification, and real-time sequencing — performed entirely in the field on reptile specimens in a tropical rainforest.⁷ A separate proof-of-concept in a Tanzanian montane forest tested the approach on a wild frog, performing the entire process on site at 26°C, 98 percent humidity, and with no electricity supply — conditions about as far from a clean lab as you can get.²⁴ Yields were low in that extreme setting, but the identification was successful. Plant species have also been identified in the field using nanopore sequencing, with the entire process from sample collection to species-level identification completed within a few hours.²⁵

Field-portable barcoding could be transformative for biodiversity surveys in remote areas, for customs inspections at ports, and for conservation work where a species identification needs to happen before the researcher leaves the site. The technology is still maturing, but it is already functional.

Where DNA Barcoding Falls Short

Barcoding is not infallible, and understanding its blind spots matters as much as appreciating what it can do.

One well-known issue is nuclear copies of mitochondrial genes, called NUMTs. Over evolutionary time, fragments of mitochondrial DNA occasionally get pasted into the nuclear genome, where they sit as non-functional relics. When PCR accidentally amplifies one of these pseudogenes instead of the real mitochondrial barcode, it introduces a misleading sequence into the analysis. In tunicates of the genus Polyclinum, NUMTs have been flagged as a genuine impediment to COI-based barcoding.²⁶ In broad marine metabarcoding surveys, NUMTs could inflate the apparent count of species by about 9 percent when the standard barcode fragment length and position are accounted for, a manageable distortion but one worth being aware of.²⁷

Hybridization is another weakness. Because standard animal barcoding reads mitochondrial DNA, which is inherited only from the mother, a hybrid animal will carry the barcode of its maternal species regardless of what its father was. Species that naturally hybridize or have a history of genetic mixing between populations can produce misleading barcode results, and this problem is often underappreciated in large automated analyses.²⁸

Reference library gaps represent a different kind of problem. If the species you are trying to identify is not yet represented in BOLD or GenBank, barcoding cannot give you a name. It can tell you the unknown specimen is distinct from everything else in the database, which is a useful signal, but you will still need a taxonomist to describe it formally. And for groups that are poorly studied or found in biodiversity hotspots where sampling has been sparse, library gaps remain significant.

What It Costs

A common assumption is that molecular methods must be cheaper and faster than hiring a human taxonomist to sort through specimens under a microscope. That assumption is only partly right. A survey of freshwater bioassessment programs in the United States found that Sanger-based barcoding was roughly 1.7 to 3.4 times more expensive than traditional morphological identification, once field-sampling costs (which are shared by both approaches) are excluded. However, next-generation sequencing methods brought the cost down to roughly comparable or slightly less expensive than morphological methods, depending on the type of organism being assessed.²⁹

The cost trajectory is clearly downward. Sequencing prices continue to fall, and the compressed reference libraries and automated bioinformatics pipelines being developed reduce the computational burden per sample. Where barcoding has a clear economic edge is in situations requiring the processing of large volumes of similar-looking specimens — bulk insect trap samples, regulatory seafood testing, customs inspections — where a trained human would need unrealistic amounts of time.

When Barcoding Meets Traditional Taxonomy

DNA barcoding was initially controversial among taxonomists. Critics saw it as a threat to the discipline, reducing the rich art of species description to reading a short gene fragment. That debate has largely settled into a working partnership. Barcoding is excellent at flagging specimens that may belong to an undescribed species or that have been misidentified, but it cannot, on its own, describe a new species. Formal species descriptions still require morphological analysis, ecological data, and often multiple genetic markers beyond the standard barcode region.

Where the two approaches reinforce each other is in biodiversity inventories. A research team can barcode thousands of specimens quickly, sort them into genetic clusters that likely represent distinct species, and then direct taxonomic expertise to the clusters that look unusual or unmatched. This triage function is arguably barcoding’s greatest contribution to taxonomy — it focuses limited expert attention where it is most needed, rather than replacing that expertise altogether.