A customs officer stands on a humid dock surrounded by crates of dried shark fins, sweat dripping down her neck as she take a small sample for testing. The air smells like salt and diesel, and in the background people are having conversations amongst themselves. She pays no mind to them, securing that tiny sliver of fin into a tube, closing a portable qPCR machine the size of a lunchbox, and tapping “Start.” Within minutes the DNA begins to amplify. Instead of waiting days for lab results, ninety minutes later the screen shows a clean melt curve, then an automated species name. Two hours from start to finish and she knows exactly which shark species she’s dealing with. If it’s on the CITES list and the shipment lacks proper permits, enforcement actions can begin immediately.
No delays.
No guesswork.
No traffickers slipping away.
Sharks and rays are some of the most threatened animals in our oceans and the demand for shark fins, driven largely by cultural cuisines like shark fin soup, is one of the many threats they face pushing their populations down worldwide. To address these threats, international trade in sharks and rays is regulated under CITES, the Convention on International Trade in Endangered Species of Wild Fauna and Flora. Some species, like at least five sawfishes from the family Pristidae, are listed under Appendix I, meaning they are threatened with extinction and trade is only allowed in exceptional circumstances. Over 100 other species fall under Appendices II and III. Appendix II species aren’t necessarily on the brink of extinction, but their trade is controlled to prevent overexploitation. Appendix III includes species that are protected in at least one country, which has asked other CITES Parties for help managing their trade.
Yet illegal trade still exists, and is one of the most stubborn challenges in wildlife conservation. Globally, these crimes have become the fourth most profitable illegal activity, worth an estimated $20 billion dollars (USD) each year. With that kind of money at stake, enforcement officers face enormous pressure to move fast and build airtight cases. Yet one of the earliest steps in stopping wildlife crime often proves to be the hardest: figuring out exactly which species they are looking at.
Many officials worldwide have been trained to identify shark fins in all states (unprocessed or dried) and TRAFFIC has even created a 3D Fin Identification Guide to help law enforcement officials identify the fins of 11 commercially traded shark and rays species based on physical characteristics (those most commonly found in international trade). But it’s still a very difficult job to do, and even experienced experts can’t reliably tell certain species apart just by sight, especially small or heavily processed pieces. So when fins are dried, meat is processed, or body parts are chopped into unrecognizable pieces, how can anyone tell which species is which?
Traditional DNA often used primers designed to detect one species at a time. If you want to check dozens of species in a single shipment, you need multiple primers, and multiplexing becomes complicated. Too many primers can interfere with PCR reactions and reduce efficiency. Not to mention that sequencing, the gold standard for precise identification, is expensive and time-consuming, often taking days to weeks. In high-volume enforcement scenarios, like checking tons of shark fins at a port, waiting for lab results is impractical. By designing a portable, low-cost version of this technique, a team of researchers led by Dr. Diego Cardeñosa of Florida International University have created a tool that frontline officers can take directly into ports, markets, or airports. It works in as little as two hours and costs only $1.50 USD per sample. And they put it to the test in a new study led by Cardeñosa, where the team highlighted high-resolution melt analysis, or HRM, a genetic method that identifies species by heating DNA until it “melts” into single strands; the temperature and shape of that melt curve are unique to each species, similar to our own, unique fingerprints except this is a molecular fingerprint.
To build the assay, researchers gathered 669 tissue samples from 66 different shark and ray species. These samples came from real-world contexts (think market surveys, field operations, meat stored in freezers, dried fins, even cooked shark fin soup) because their goal was not to create a tool that was only good when samples were “pristine”or in a perfect scenario. They want technology that can withstand the messy, unpredictable samples that enforcement officers encounter daily. Using a small region of the mitochondrial 12S rRNA gene, they designed universal primers that amplify a short but highly variable sequence. Those little differences between species may not seem like a big deal, but they translate into distinct melt curve shapes once the DNA is heated. And the end results were impressive: the assay successfully distinguished at least 55 shark and ray species, including 38 listed under CITES! Enforcement officers routinely intercept shipments with fins from silky sharks (Carcharhinus falciformis), scalloped hammerheads (Sphyrna lewini), oceanic whitetips (Carcharhinus longimanus), and other threatened species, yet these products can be so processed that even experts cannot identify them by sight. If a customs officer could instead run a sample through a suitcase-sized qPCR machine, wait two hours, and receive an automated classification, what would that mean for real-time decision making?
The machine learning models also further strengthened the process by automatically recognizing melt curve images. For example, a lightweight ResNet18 model achieved 99.2 percent accuracy (which matters when an entire shipment hinges on whether a species is correctly identified), outperforming more complex algorithms. The system made sure to include confidence thresholds as well, so officers know when a result is uncertain and requires follow-up sequencing. What makes this tool compelling is not only its accuracy (and ability to state when it isn’t confident about an identification), but its reliability with difficult samples; fins that were dried, salted, cooked, or collected individually from soup still produced identifiable melt profiles! Even when confidence values dropped because the DNA quality was poor, species assignments still remained correct. This suggests the method is robust enough to handle even the most challenging “materials” that traffickers often rely on to hide their tracks. But this method isn’t perfect and there were some limitations, particularly among closely related species. Members of the Carcharhinus genus can have highly similar melt profiles due to only a handful of nucleotide differences. For example, dusky sharks (Carcharhinus obscurus) and oceanic whitetips occasionally produced overlapping signatures. Yet in most enforcement scenarios, identifying that a shipment contains a CITES-listed species within that complex — rather than a specific species — is enough to justify further investigation. Which then begs the question: “Does enforcement always need perfect taxonomic precision, or is rapid, actionable information more valuable on the ground?”
This research also makes one question a broader question about conservation technology. As genetic tools become cheaper and more portable, how can they be integrated into global policy frameworks? CITES listings require countries to prove that traded wildlife products are legally and sustainably sourced, but without fast species identification, these regulations become difficult to enforce. Cardeñosa and his team are making the possibility of a world where customs officers regularly carry portable DNA testers, much like breathalyzers at roadside checks, a reality. Could that shift the balance of power away from traffickers who rely on anonymity? The new technology also has implications for how we think about transparency in global trade: if two hours and $1.50 USD can provide species-level identification at a border, what other industries might adopt similar rapid genetic tools? Could fisheries audits, seafood fraud investigations, or even ecological monitoring become faster and more equitable by placing molecular tools directly in the hands of local authorities?
The HRM assay is not the final word in species identification. Sequencing will always be needed for borderline cases, reference libraries must grow to capture global genetic diversity, and machine learning models must keep learning as new species are added. Yet this study presents one of the most comprehensive and scalable approaches to shark and ray identification ever developed.
In a world where wildlife trafficking adapts quickly, conservation science must innovate just as fast. This new HRM assay is one heck of a good start.
