On-site metabarcoding analysis of environmental DNA samples

This study demonstrates the feasibility of a fully portable, off-grid eDNA metabarcoding pipeline using compact equipment like the BentoLab and MinION sequencer, which successfully enabled on-site collection, sequencing, and analysis of aquatic biodiversity in remote locations within a few days.

The Gist

Imagine you want to know what fish are swimming in a river, but you can't see them, catch them, or even get close to them without scaring them away. Traditionally, scientists would have to take a bucket of water, drive it back to a high-tech lab, wait weeks for results, and hope the DNA in the water didn't rot on the way.

This paper is about a team of scientists who decided to turn that process on its head. They built a "mobile DNA detective kit" that fits in a backpack and can solve the mystery of "who's in the water" right there on the riverbank, all within a few days.

Here is the story of how they did it, explained simply:

1. The Problem: The "Wait-and-See" Game

Usually, environmental DNA (eDNA) work is like sending a letter to the post office and waiting months for a reply. You collect water, ship it to a lab, wait for the DNA to be read, and by the time you get the answer, the fish might have moved, or the water might have changed. It's slow, expensive, and requires a fancy building full of expensive machines.

2. The Solution: The "Backpack Lab"

The researchers wanted to prove you don't need a palace to do science; you just need a toolkit. They packed everything into a portable setup:

• The BentoLab: Think of this as a "Swiss Army Knife" for biology. It's a small box that can spin samples (like a salad spinner) and heat them up to cook the DNA, all on a battery.
• The MinION Sequencer: This is the star of the show. It's a USB stick-sized device that reads DNA. Imagine it as a tiny, portable barcode scanner that can read the genetic "license plates" of fish without needing a massive supercomputer.

3. The Mission: Two Training Camps

The team didn't just test this in a quiet lab; they took it to the field in two places:

• Brazil: A tropical classroom where students learned to fish for data.
• Norway: A cold, coastal classroom where they tested the gear in the Oslofjord.

The goal wasn't just to catch fish; it was to teach the next generation of scientists how to be "field detectives" who can get answers immediately.

4. How It Worked: The "Water Soup" Recipe

Here is the step-by-step process they used, which they completed entirely on-site:

• Step 1: The Scoop. They dipped a sterile bag into the water (like scooping soup) and filtered it through a tiny coffee filter. This caught the invisible "skin cells" and "poop" (DNA) that fish leave behind in the water.
• Step 2: The Extraction. They put the filter in the BentoLab, added some chemical "soup," and heated it up. This broke open the cells to release the DNA, like squeezing a grape to get the juice.
• Step 3: The Amplification. They used a "molecular photocopier" (PCR) to make billions of copies of the fish DNA. Since there was so little DNA in the water, they needed to make it loud enough to hear.
• Step 4: The Reading. They loaded the DNA onto the MinION (the USB stick). Over the course of about 18 hours, the device read the genetic code, turning the water sample into a list of names.
• Step 5: The Translation. Using a laptop (no internet needed!), they ran a program that translated the genetic code into fish names, like a translator turning a foreign language into English.

5. The Results: A Successful Field Test

In the Oslofjord, they found 16 different species of fish and sharks, including some rare ones like the Atlantic halibut and the thornback ray.

• The "Aha!" Moment: They compared their results to a list of fish known to be in the aquarium at the Drøbak Marine Station. The portable kit found all the fish in the tank, proving it worked.
• The Catch: They found fewer species than big lab studies usually find (which is expected because they took fewer samples and used a slightly different "scanner"). However, the fact that they got any reliable results in the field, without a lab, was the real victory.

6. Why This Matters: The "Fire Alarm" Analogy

Think of traditional lab testing like checking the weather report from last week. It tells you it rained, but it doesn't help you if you need an umbrella now.

This new portable method is like a real-time fire alarm.

• If a toxic algae bloom starts, you can test the water today and warn the town today.
• If an invasive species (a "bad neighbor" fish) shows up, you can spot it immediately and stop it before it takes over.
• It brings science to remote places (like the Amazon or the Arctic) where there are no labs, empowering local people to protect their own waters.

The Bottom Line

This paper proves that high-tech science doesn't have to be stuck in a concrete building. By shrinking the lab down to the size of a suitcase, the team showed that we can monitor the health of our oceans and rivers faster, cheaper, and right where the action is. It's a giant leap toward making biodiversity monitoring as easy as taking a photo.

Technical Summary

1. Problem Statement

Environmental DNA (eDNA) metabarcoding is a powerful tool for monitoring aquatic biodiversity, fisheries, and aquaculture impacts. However, conventional workflows face significant limitations:

• Logistical Constraints: They rely on dedicated laboratory infrastructure, specialized equipment, and long turnaround times for sample transport and processing.
• Sample Degradation: Delays between collection and processing can lead to DNA degradation, particularly in remote locations.
• Lack of Real-time Feedback: The inability to process samples on-site prevents researchers from adapting sampling strategies (e.g., location, intensity, or primer selection) in real-time based on initial findings.
• Educational Gaps: There is a need to train the next generation of scientists in rapid, portable molecular techniques outside of traditional clean labs.

2. Methodology

The study presents a fully portable, off-grid eDNA metabarcoding pipeline tested during two international training courses (in Norway and Brazil). The workflow was executed entirely on-site using compact equipment.

• Equipment:
  • BentoLab: A portable all-in-one workstation (centrifuge, thermal cycler, electrophoresis) used for DNA extraction and PCR.
  • Oxford Nanopore Technologies (ONT) MinION: A portable sequencer used for library preparation and sequencing.
  • Standard Laptop: Used for bioinformatics without internet access.
• Sampling:
  • Location: Oslofjord, Drøbak, Norway (Nov 3–4, 2025).
  • Samples: 7 water samples (6 from the Oslofjord, 1 from the Drøbak Aquarium as a positive control).
  • Volume: 750 mL to 2500 mL filtered through 0.8 μm Whatman filters.
• Molecular Workflow:
  • Extraction: Qiagen DNeasy Blood & Tissue Kit (modified protocol with doubled reagent volumes) performed in the BentoLab.
  • Amplification: Two primer sets targeting the mitochondrial 12S rRNA gene were used:
    • Riaz primers: 106 bp fragment.
    • Teleo2 primers: ~167 bp fragment.
  • Replication: Each sample underwent four PCR replicates.
  • Library Prep: Modified SQK-NBD114-24 protocol. Indexed primers were used for initial tagging, followed by ONT Native Barcodes (NB01–NB04) for multiplexing.
  • Sequencing: 18 hours 40 minutes on a FLO-MIN114 flow cell using the High-accuracy-v4.3.0 model (Q-score ≥ 9).
• Bioinformatics:
  • Processing: Performed offline using cutadapt (demultiplexing), VSEARCH (quality filtering, dereplication, chimera removal, OTU clustering at 97%), and SINTAX/usearch_global for taxonomic assignment against the MIDORI2 database.
  • Validation: Taxonomic assignments were cross-referenced with local occurrence records (Artsdatabanken, GBIF) and corrected via BLAST against NCBI to remove false positives caused by short amplicon lengths.
  • Statistical Analysis: Conducted in R (packages: vegan, ggplot2, dplyr) for diversity metrics, accumulation curves, and community composition analysis.

3. Key Contributions

• Proof of Concept for Off-Grid Analysis: Demonstrated that a complete eDNA metabarcoding workflow—from water filtration to final species list—can be completed in five days in a field setting without access to a clean lab or high-performance computing clusters.
• Portable Training Model: Successfully trained students and early-career researchers from diverse backgrounds (Brazil, Norway, South Africa) in a portable, hands-on environment, fostering international collaboration and technical competence.
• Open-Source Protocol: Provided a fully reproducible protocol, including custom R scripts and bioinformatics pipelines, available via Zenodo, specifically designed for resource-limited settings.
• Primer Bias Analysis: Highlighted the critical impact of primer selection on detection capabilities by comparing Riaz/Teleo2 results against previous Illumina-based studies in the same region.

4. Results

• Sequencing Output: Generated 13.43 million reads (~8.04 Gb). After filtering, 3.09 million reads were retained, representing 1,053 unique OTUs.
• Species Detection:
  • Identified 16 fish and elasmobranch species across the samples.
  • Positive Control: The aquarium sample successfully detected all 8 species physically present in the tank, plus 8 additional species likely introduced via seawater intake.
  • Field Samples: Detected key species including Anarhichas denticulatus (northern wolffish), Anarhichas lupus (Atlantic wolffish), Hippoglossus hippoglossus (Atlantic halibut), Raja clavata (thornback ray), and Scyliorhinus canicula (small-spotted catshark).
• Taxonomic Corrections: Six OTUs initially misidentified by the MIDORI2 database (e.g., Dipturus batis instead of Raja clavata) were corrected using local occurrence data and BLAST, underscoring the need for local validation.
• Comparison with Previous Studies:
  • Lower Diversity: The study detected only 16 species compared to 51–65 species in previous Oslofjord studies using Illumina sequencing.
  • Reasons for Discrepancy: Attributed to the limited sampling volume, single location, short duration, and the higher error rate of Nanopore sequencing requiring stringent filtering (which may have removed rare true positives).
  • Primer Influence: The Riaz and Teleo2 primers detected species (e.g., wolffish, halibut) that were missed by MiFish and Elas02 primers used in previous studies, demonstrating that primer choice significantly influences community composition results.
• Community Composition: No significant differences were found in community composition between the two primer sets or between the two sampling dates.

5. Significance

• Rapid Response Capability: The protocol enables near-real-time ecological assessments, crucial for managing invasive species, monitoring aquaculture pathogens, or responding to environmental disasters in remote areas.
• Democratization of Science: By removing the dependency on expensive, centralized infrastructure, this approach makes advanced biodiversity monitoring accessible to researchers in developing nations and remote field stations.
• Educational Impact: The successful execution of the workflow during training courses proves that complex molecular techniques can be taught effectively outside traditional university laboratories.
• Future Applications: While the study noted lower sensitivity compared to Illumina, the ability to generate actionable data within days makes this workflow ideal for rapid decision-making in fisheries management and conservation, provided that limitations regarding primer bias and sequencing depth are accounted for.