A conservation planning assessment of basin wide Unionid mussel assemblages using environmental DNA

This study demonstrates that environmental DNA metabarcoding is a sensitive and efficient alternative to traditional visual surveys for assessing basin-wide unionid mussel assemblages, successfully detecting a higher number of species—including rare and federally protected ones—while offering valuable insights into occupancy patterns when detection repeatability is considered.

The Gist

Imagine you are trying to find a specific type of rare, shy animal living in a river. These animals are freshwater mussels. They are like the "ghosts" of the river world: they spend most of their time buried deep in the mud, sand, or under rocks, and they rarely come out to say hello.

For decades, scientists have tried to count them by getting into the water, feeling around with their hands, and turning over rocks. This is like trying to find a needle in a haystack while wearing thick winter gloves. It takes a long time, it's exhausting, and you often miss the ones hiding the best.

This paper is about a new, high-tech way to find these "ghosts" using Environmental DNA (eDNA).

The "Scent Trail" Analogy

Think of eDNA like a scent trail left behind by a dog. Even if the dog isn't in the room, it leaves behind tiny bits of fur, skin cells, and waste in the air. If you have a super-sensitive nose (or in this case, a DNA lab), you can sniff the air and say, "A dog was definitely here," even if you never saw the animal.

Freshwater mussels do the same thing. They constantly shed tiny bits of DNA into the water through their waste and skin. Scientists can scoop up a cup of river water, filter out these microscopic DNA "breadcrumbs," and run them through a machine that reads the genetic code. This tells them exactly which species of mussels are swimming (or burrowing) in that stretch of river.

What the Scientists Did

The researchers took this new "scent-sniffing" method and compared it to the old "rock-turning" method in Fish Creek, a river that flows through Ohio and Indiana.

1. The Old Way (Visual Surveys): Teams of biologists waded into the river for hours, turning over rocks and digging in the mud. They found 22 different species of live mussels.
2. The New Way (eDNA): They took water samples from the same spots and analyzed the DNA. They found 25 different species.

The Big Wins

The new method was like upgrading from a flashlight to a night-vision camera. Here is what they discovered:

• Finding the Elusive: The eDNA method found four species that the human divers completely missed. One of these was the Salamander Mussel, a rare species that is notoriously hard to find because it hides under big, flat rocks or in the river banks. The DNA method "smelled" it, leading scientists to go back and find it with their own eyes later.
• The "Ghost" Check: They were specifically looking for a rare mussel called the "White Cat's Paw," which hadn't been seen since 1999. Neither the divers nor the DNA found it. This suggests the sad news that this species might actually be gone from this river.
• The "Scent" Distance: The paper explains that eDNA is so sensitive it can detect mussels that are actually living upstream or in a different part of the river, because the water carries their DNA downstream. It's like smelling a campfire from a mile away; you know the fire is there, even if you can't see the flames yet.

How They Made Sense of the Data

Because eDNA is so sensitive, it sometimes picks up DNA from mussels that aren't right at that exact spot (just like smelling a neighbor's BBQ). To fix this, the scientists created a "Repeatability Score":

• High Repeatability: If the DNA showed up in almost every water sample taken at a spot, they knew the mussels were definitely living right there. This matched perfectly with what the divers saw.
• Low Repeatability: If the DNA showed up only once or twice, it likely meant the mussels were nearby (upstream or downstream) and their DNA drifted by.

Why This Matters

This study proves that eDNA is a super-efficient tool for conservation.

• Speed: Collecting water samples takes about 40 minutes. Digging for mussels takes hours.
• Sensitivity: It finds rare species that humans miss.
• Cost: It's cheaper and less disruptive to the river habitat.

The Bottom Line

The authors aren't saying we should stop looking at the mussels with our eyes. Instead, they suggest a teamwork approach:

1. Use the eDNA "scent sniff" first to quickly scan a large area and find where the rare species might be.
2. Send the divers in for a targeted, intensive search in those specific spots to confirm the population.

It's like using a metal detector to find where to dig for treasure, rather than digging randomly across the whole beach. This combination gives conservationists the best chance to protect these imperiled creatures before they disappear forever.

Technical Summary

1. Problem Statement

Freshwater mussels (Order Unionida) are among the most imperiled faunal groups in North America, with over 75 federally endangered and 21 threatened species in the U.S. Traditional survey methods for these organisms are labor-intensive, costly, and heavily reliant on expert taxonomists. Furthermore, mussels often burrow into benthic substrates or exhibit cryptic behaviors, leading to underrepresentation in traditional surveys. This lack of comprehensive distribution and abundance data hinders effective conservation planning and recovery efforts for rare species. There is a critical need for innovative, efficient, and sensitive monitoring tools to assess mussel assemblages across large spatial scales.

2. Methodology

The study was conducted over 30 km of Fish Creek (Ohio and Indiana, USA) during the late summers of 2021, 2022, and 2023. It employed a comparative approach using two distinct survey methods:

• Visual Surveys:

  • Protocol: Low-effort qualitative surveys designed to maximize spatial coverage rather than exhaustive site-specific counts. Biologists searched riffle sections (approx. 25m) by moving upstream, disturbing the upper 5 cm of substrate, and using viewing buckets where water clarity permitted.
  • Effort: Search areas varied (38–330 m²), and search times ranged from 1.67 to 6.67 hours per site. Total effort was recorded as Catch Per Unit Effort (CPUE).
  • Scope: 39 total sites surveyed across three years (13 in 2021, 14 in 2022, 12 in 2023).

• Environmental DNA (eDNA) Surveys:

  • Sampling: Conducted prior to visual surveys to avoid sediment disturbance.
    • 2021: Three 500 mL point samples across the stream width.
    • 2022–2023: 500–1000 mL samples collected via peristaltic pump along a transect.
  • Processing: Water was filtered through 1.2 µm glass microfiber filters. DNA was extracted using a modified Qiagen DNeasy protocol, followed by inhibitor removal.
  • Molecular Analysis: Targeted a ~175 bp fragment of the mitochondrial 16S gene. Samples underwent triplicate PCR amplification, Illumina MiSeq metabarcoding (2×300 nt), and bioinformatic processing using QIIME 2 and DADA2.
  • Bioinformatics: Sequences were clustered into Molecular Operational Taxonomic Units (MOTUs) at 97.5% similarity. Taxonomic assignment was performed via BLAST against a custom database and NCBI GenBank.
  • Quality Control: Strict controls included field blanks, filtration negatives, and lab negatives. "Mistags" (index hopping) were estimated and removed using a conservative rate of 0.0075.
  • Detection Index: An eDNA Detection Index was developed based on repeatability across replicates:
    • High: Detected in ≥75% of replicates.
    • Moderate: Detected in 50–74% of replicates.
    • Low: Detected in <50% of replicates.

• Statistical Analysis:

  • Logistic regression and Bayesian integrated occupancy models (using the spOccupancy R package) were used to estimate site occupancy ($\psi$) and detection probability ($p$).
  • Covariates included distance from the St. Joseph River confluence (proxy for drainage area and habitat).
  • Correlations were tested between eDNA read counts and visual CPUE.

3. Key Contributions

• Validation of eDNA for Unionids: Demonstrated that eDNA metabarcoding is a viable, high-sensitivity tool for assessing complex freshwater mussel assemblages, detecting species missed by traditional visual surveys.
• Repeatability Framework: Introduced a "detection repeatability" metric to distinguish between local site occupancy (high repeatability) and reach-scale occupancy or transport artifacts (low/moderate repeatability).
• Discovery of Rare Species: Successfully detected Simpsonaias ambigua (Salamander Mussel), a species rarely encountered in conventional surveys and difficult to locate due to its habit of burrowing under large boulders or in bedrock crevices.
• Occupancy Modeling: Developed robust occupancy models that correlated eDNA detection probabilities with visual abundance, validating the use of eDNA for estimating species distribution at a basin scale.

4. Key Results

• Species Richness:

  • Visual Surveys: Detected 22 live species.
  • eDNA Surveys: Detected 25 species.
  • Discrepancies: eDNA detected four species not observed alive visually: Lampsilis fasciola, Simpsonaias ambigua, Toxolasma parvum, and Utterbackia imbecillis. Conversely, visual surveys missed Ortmanniana ligamentina (detected only once visually).
  • Federally Protected Species: Both methods confirmed the presence of three federally protected species (Paetulunio fabalis, Pleurobema clava, Theliderma cylindrica). eDNA was the sole method to detect S. ambigua.
  • Extirpation: Neither method detected Epioblasma perobliqua (White Cat's Paw), suggesting potential extirpation from Fish Creek.

• Detection Correlation:

  • Congruence: 94.7% of visual observations were matched by eDNA.
  • Repeatability: High-repeatability eDNA detections strongly aligned with visual presence (80.1% of visual observations). Moderate and low repeatability detections often occurred at sites without visual confirmation, likely representing downstream transport or broader reach-scale occupancy.
  • Abundance Correlation: There was a strong positive correlation between eDNA sequence read counts and visual CPUE ($R^2 = 0.63$, $p < 0.001$).

• Efficiency and Sampling Design:

  • Field Time: eDNA sampling took ~40 minutes per site (3 replicates) vs. ~4.5 hours for visual surveys.
  • Detection Probability: For the rarest observed species (CPUE 0.16), 4 samples were required to reach a 0.75 detection probability, while 9 samples were needed for 0.95 probability.
  • Occupancy: Occupancy models showed high congruence between eDNA estimates and visual distributions, particularly when filtering for high-repeatability detections.

5. Significance

This study provides compelling evidence that eDNA metabarcoding can serve as a primary or complementary tool for freshwater mussel conservation planning.

• Enhanced Sensitivity: eDNA outperformed visual surveys in detecting rare, cryptic, or low-abundance species, including S. ambigua, which is critical for recovery planning.
• Cost-Effectiveness: While laboratory processing adds time, the field efficiency of eDNA allows for broader spatial coverage at a fraction of the labor cost of traditional surveys.
• Strategic Application: The authors propose a phased sampling approach: using eDNA for broad-scale screening to identify potential hotspots for rare species, followed by targeted, intensive visual surveys for ground-truthing.
• Limitations & Context: The study highlights that eDNA signals can be influenced by hydrological transport (detecting species upstream of the sampling point) and that rare species may require higher sampling rigor (more replicates) to achieve high detection confidence.

In conclusion, integrating eDNA with species ecology and hydrological context offers a powerful, scalable solution for monitoring imperiled freshwater mussel assemblages, significantly enhancing the ability of agencies like the U.S. Fish and Wildlife Service to assess status and plan recovery.