Longer DNA marker proves reliable for routine diatom river monitoring

A quiet revolution has been unfolding in the way Europe monitors the health of its rivers. Instead of peering through microscopes to count and identify diatoms, the microscopic algae whose cell walls record the pollution history of a stream, environmental agencies are increasingly turning to DNA metabarcoding. A new study published in the journal Metabarcoding and Metagenomics now offers a decisive answer to one of the method’s most stubborn questions: which genetic marker should laboratories use for routine biomonitoring? The verdict, drawn from real river samples collected across the Czech Republic, is that the longer of two widely used markers delivers better species resolution without compromising the ecological verdicts that regulators depend on.

Diatoms are the workhorses of freshwater ecological assessment. Because different species thrive under different nutrient, pH, and pollution conditions, the composition of a diatom community acts as a biological fingerprint of water quality. Under the European Water Framework Directive, this fingerprint has traditionally been read through painstaking microscopy, with specialists identifying hundreds of cells per sample and calculating indices such as the Biological Diatom Index, or IPS. DNA metabarcoding promises to automate much of this work, offering higher throughput and better reproducibility, but its adoption hinges on a critical methodological choice: the DNA barcode region selected for amplification.

The chloroplast gene rbcL, which encodes part of the enzyme responsible for carbon fixation during photosynthesis, has emerged as the most promising target for diatom metabarcoding. Yet different regions of this gene, varying in length and position, have been explored in different studies, ranging from short fragments suitable for degraded sedimentary DNA to longer amplicons that generally provide higher taxonomic resolution. Two markers dominate current practice. The shorter one amplifies a 263-base-pair fragment using a highly degenerate multiplex primer set, while the longer one amplifies a 331-base-pair fragment with a single primer pair of lower degeneracy. The entire 263-base-pair region sits within the longer amplicon, corresponding to its three-prime end, with the longer marker extending an additional 68 bases toward the five-prime end.

These differences matter more than they might appear. Reduced marker length can limit the resolution of closely related taxa and mask intraspecific variation, while differences in primer design can introduce amplification biases that affect the detection and relative abundance of species. Until now, it remained unclear how such differences propagate through the entire analytical pipeline, from raw sequence data to species composition and, crucially, to abundance-weighted ecological indices like IPS under realistic biomonitoring conditions. A team of researchers set out to answer this question directly by analysing identical benthic diatom samples with both primer sets, testing whether marker choice affects ecological conclusions or whether abundance-weighted indices remain robust across markers.

The study drew on eighteen epilithic biofilm samples collected in 2024 during routine river biomonitoring in the Ohře River basin, conducted by the Czech state enterprise Povodí Ohře. Sampling followed the standardized biofilm collection protocol CEN/TR 17245, with negative field controls processed alongside environmental samples. The sites spanned a pronounced environmental gradient, particularly in nutrient concentrations, with nitrate-nitrogen ranging from 0.52 to 5.20 milligrams per litre and phosphate-phosphorus from 0.02 to 0.18 milligrams per litre, along with conductivity spanning 11.8 to 105.7 millisiemens per metre and suspended solids from 2.5 to 53.8 milligrams per litre. Samples were fixed in 96 percent ethanol immediately after collection and stored at minus twenty degrees Celsius until DNA extraction, which was performed in duplicate using a modified CTAB protocol with negative extraction controls included in each batch.

Sequencing on an Illumina MiSeq platform yielded a median of roughly 206,000 reads per sample for the shorter marker and 196,000 for the longer one. After quality filtering and chimera removal, median retention rates stood at 84 percent and 74 percent respectively. Off-target amplification proved negligible for both markers, with off-target green lineages accounting for less than a third of one percent of raw reads. Technical replicates showed high consistency, with concordance increasing with sequence variant abundance and approaching saturation for the most abundant variants. The researchers then processed the raw data through the DADA2 pipeline, inferring amplicon sequence variants with error models estimated separately for each sequencing run, and normalised all datasets to relative abundances.

At the sequence level, the two markers told a strikingly consistent story. Of the roughly 600 red-lineage sequence variants detected in each dataset, about 70 percent were shared based on identical core sequences in the overlapping region. These shared variants accounted for the overwhelming majority of reads, 98.77 percent in the longer-marker dataset and 99.35 percent in the shorter-marker dataset. Relative abundances of shared variants correlated strongly across samples, with Spearman coefficients ranging from 0.65 to 0.87 and a median of 0.77, and overall community composition proved highly consistent between datasets, with a Mantel test correlation of 0.90. Dataset-specific variants were generally rare and occurred in few samples, and analysis of primer binding sites revealed no significant differences in mismatch number or positional distribution between shared and marker-specific variants, indicating no systematic primer bias explaining their specificity.

When the researchers moved from raw sequences to species-level identification, using the Diat.Barcode v15.2 reference database and correcting abundances for species-specific differences in rbcL copy number, the picture remained remarkably coherent. The two datasets detected similar numbers of taxa, 185 for the longer marker and 181 for the shorter one, with 159 taxa shared. After accounting for differences in sequence length, the number of genuinely marker-specific taxa shrank to just 21 and 17 respectively, all of consistently low abundance and prevalence. Species composition differed statistically between datasets, but marker identity explained only a small proportion of the variation, and paired samples were consistently more similar to each other than to any other samples in the study.

The most consequential finding concerned the IPS index itself. IPS values correlated strongly between markers, with a Pearson coefficient of 0.94, though the shorter marker tended to yield slightly higher scores, with a median difference of 0.890 points on the 20-point scale. This shift was enough to change the assigned ecological status class in three of the eighteen samples. To understand where these differences came from, the researchers performed a sequential decomposition, recalculating the index under four conditions: the full comparison, a length-harmonised comparison, a comparison excluding marker-specific taxa, and a comparison restricted to taxa shared within each sample. The results showed that most of the difference was already captured when only relative abundance differences among shared taxa were considered, with a median effect of 0.568 index points. Adding marker-specific taxa caused only negligible shifts, while annotation bias related to amplicon length produced a larger secondary contribution of 0.334 points.

The practical implications are clear. The longer 331-base-pair marker, with its higher taxonomic resolution for abundant taxa, appears well suited for routine biomonitoring under the Water Framework Directive, and its lower detection of rare taxa does not compromise ecological assessment. Rare taxa are generally not critical for index calculation, though studies focused on the rare biosphere might still benefit from multiplex primer strategies like those used for the shorter marker. The authors also note that deriving a marker-specific correction factor for the longer region could further improve agreement between markers and with morphology-based assessments, since the existing correction factor was originally established for the shorter amplicon. Taken together, the findings suggest that laboratories can adopt the longer marker in practice without major methodological adjustments, a conclusion that could accelerate the integration of DNA-based monitoring into European regulatory frameworks and bring the promise of faster, more reproducible river assessment one significant step closer to everyday reality.

Subject of Research: Comparison of two rbcL DNA markers for diatom metabarcoding in river biomonitoring

Article Title: Evaluation of two overlapping rbcL amplicons for routine diatom DNA biomonitoring

Article References: Evaluation of two overlapping rbcL amplicons for routine diatom DNA biomonitoring. (n.d.). https://doi.org/10.3897/mbmg.10.183921

Image Credits: AI Generated

DOI: 10.3897/mbmg.10.183921

Keywords: diatoms, DNA metabarcoding, rbcL marker, Water Framework Directive, biomonitoring, IPS index, ecological status, freshwater assessment, Ohře River, amplicon comparison, Diat.Barcode, river health