Researchers collected two replicate kick samples at each site using a 250-micrometre D-net. One replicate from each site was assessed morphologically, while the other was homogenized for bulk-tissue metabarcoding. DNA was extracted, amplified with the freshwater macroinvertebrate primer pair fwhF2/EPTDr2n, and sequenced with paired-end Illumina MiSeq sequencing. The molecular analysis produced 127,707 paired-end reads across the 10 samples, of which 119,320 were retained after trimming.

Morphological examination identified 50 families and 60 genera across the sites. By contrast, the metabarcoding results contained 341 molecular operational taxonomic units, or mOTUs. Of these, 328 could be assigned to order, 282 to family, 88 to genus and 22 to species. Overall, the molecular method recovered 37 families, 40 genera and 22 species after terrestrial taxa detected at five sites were removed for comparison with the aquatic macroinvertebrates identified morphologically.

The methods detected mostly different taxa. At the family level, overlap ranged from one family at sites SU-2 and PA-2 to eight at SU-1 and TU-2. Chironomidae was the only family detected by both approaches at every site. The two approaches identified at least two genera in common at half of the sites. Across all mOTUs, 17% remained unclassified at family level and 74% at genus level, indicating that much of the molecular signal could not be matched confidently to the available references.

Morphological identification produced more unique family and genus determinations per site than metabarcoding. PA-3 and TU-1 had equal family counts between the methods, while SU-3 had more genera detected by metabarcoding. The molecular approach nevertheless assigned some taxa to genus or species despite their being difficult or impossible to identify morphologically, including groups such as chironomids and culicids that could often be determined only to subfamily or tribe under the morphological protocol. Earlier instars and damaged specimens also posed challenges for microscopy that the molecular method did not encounter.

The researchers compared the methods using the Andean-Amazonian Biotic Index, or AAMBI, which assigns pollution-tolerance scores to families and sums them for each site. The two methods produced the same biological-quality category at only three sites. At the other seven, the category estimated from metabarcoding was lower than the category estimated from morphology. Despite these differences, site-level scores were correlated across methods, with Pearson’s r equal to 0.88 and p equal to 0.00077, and Spearman’s rho equal to 0.82 with p equal to 0.0037.

Several limitations constrain the comparison. Each site was represented by one sample for each method, and the researchers conducted neither extraction nor PCR replicates. The molecular analysis used a single primer pair that had previously been reported to underrepresent Trichoptera, Turbellaria and several crustaceans. Sequencing depth was also limited: rarefaction curves approached an asymptote, but the authors stated that more than 100,000 reads per sample would likely be needed to recover additional unique and rare taxa.

Reference-database coverage was a major limitation. Among the families and genera identified morphologically, 26 families and 40 genera had no representative sequences from Ecuador in BOLD or GenBank when checked in October 2023. Families including Ceratopogonidae, Empididae and Limoniidae, and genera including Heterelmis, Tenagobia, Eupera and Psephenus, were among those without local sequences. The incomplete representation of Ecuadorian taxa limited the number of mOTUs that could be assigned beyond class or order and reduced overlap between the methods.

The findings show that bulk-tissue metabarcoding produced biological-quality index scores that were significantly associated with those from morphology, despite detecting fewer named taxa and a different taxonomic composition at most sites. The study involved only 10 Ecuadorian locations, one field replicate per method, one primer pair and a single sequencing run, so it does not establish performance across all Neotropical freshwater systems. The authors concluded that DNA identification can complement traditional monitoring but, with current reference databases and primer performance, cannot replace morphological identification in Neotropical macroinvertebrate biomonitoring.

Subject of Research: Comparison of morphological and DNA metabarcoding methods for monitoring macroinvertebrate communities in 10 Ecuadorian streams and rivers

Article Title: Comparing morphological and DNA-based bioassessment methodologies for macroinvertebrates in Neotropical streams: a case study from Ecuador

Article References: Jijón, G., Errigo, I. M., Wicks, J., Goldston, N. N., Davis, L., Davis, D., Standring, S., Chaston, J. M., Frandsen, P. B., & Rios-Touma, B. (2025). Comparing morphological and DNA-based bioassessment methodologies for macroinvertebrates in Neotropical streams: a case study from Ecuador. Metabarcoding and Metagenomics, 9, Article e138172. https://doi.org/10.3897/mbmg.9.138172

Image Credits: AI Generated

DOI: 10.3897/mbmg.9.138172

Keywords: DNA metabarcoding, macroinvertebrates, freshwater biomonitoring, morphological identification, Ecuador, Neotropical streams, AAMBI, DNA barcoding, reference databases, Comparing, morphological, DNA-based