Hidden diversity – DNA metabarcoding reveals hyper-diverse benthic invertebrate communities

Abstract

Freshwater ecosystems, such as streams, are facing increasing pressures from agricultural land use. Aquatic insects and other macroinvertebrates have historically been used as indicators of ecological condition and water quality in freshwater biomonitoring programs; however, many of these protocols use coarse taxonomic resolution (e.g., family) when identifying macroinvertebrates. The use of family-level identification can mask species-level diversity, as well as patterns in community composition in response to environmental variables. Recent literature stresses the importance of robust biomonitoring to detect trends in insect decline globally, though most of these studies are carried out in terrestrial habitats. Here, we incorporate molecular identification (DNA metabarcoding) into a stream biomonitoring sampling design to explore the diversity and variability of aquatic macroinvertebrate communities at small spatial scales. We sampled twenty southern Ontario streams in an agricultural landscape for aquatic macroinvertebrates and, using DNA metabarcoding, revealed incredibly rich benthic communities which were largely comprised of rare taxa detected only once per stream despite multiple biological replicates. In addition to numerous rare taxa, our species pool estimates indicated that after 240 samples from twenty streams, there was a large proportion of taxa present which remained undetected by our sampling regime. When comparing different levels of taxonomic resolution, we observed that using OTUs revealed over ten times more taxa than family-level identification. A single insect family, the Chironomidae, contained over one third of the total number of OTUs detected in our study. Within-stream dissimilarity estimates were consistently high for all taxonomic groups (invertebrate families, invertebrate OTUs, chironomid OTUs), indicating stream communities are very dissimilar at small spatial scales. While we predicted that increased land use would homogenize benthic communities, this was not supported as within-stream dissimilarity was unrelated to land use.

Metabarcoding metacommunities: time, space, and land use interact to structure aquatic macroinvertebrate communities in streams

“…Streams are very heterogenous habitats and support many rare species, which can make it challenging to describe patterns in invertebrate distributions (Heino et al, 2015). Previous research using DNA metabarcoding in Ontario streams supported this description of heterogenous streams as aquatic macroinvertebrate communities were highly variable within a single stretch and largely composed of rare OTUs (Gleason et al 2022). However, while Heino et al (2015) detected only weak relationships between environmental and spatial factors and macroinvertebrate community structure at genus-level resolution, we found significant relationships between all variables (local, regional, spatial) even though DNA metabarcoding increases the number of rare taxa detected.…”

“…However, Noguerales et al (2021) found that both environmental filtering and spatial arrangement were important structuring components of soil invertebrates. These studies illustrate the applicability of DNA metabarcoding for metacommunity ecology, and a key uncertainty which remains is how metacommunity dynamics change in response to agriculture in stream systems where 1) there are no discrete community boundaries and 2) community composition can vary drastically at small spatial scales (Gleason et al, 2022).…”

“…No sample was flagged as containing contaminants, but there was a small number of reads in the sequencing controls, indicating tag-jumps had likely occurred during sequencing. To control for the influence of tag-jumps in the dataset, we selected a 0.001% abundance filter threshold (i.e., an OTU representing a relative abundance of less than 0.001% of total sequences of that OTU was converted to a zero in affected samples; see Gleason et al 2022; Zinger et al, 2021). We only included benthic macroinvertebrate taxa in our dataset (arthropods, molluscs, annelids) and required a minimum match similarity threshold of 90% to the reference database (BOLD).…”

“…In a call to action for freshwater preservation, Maasri et al (2022) identified multiple research priorities, including ecology and biomonitoring, and suggest that these fields can benefit from the incorporation of molecular methods for taxonomic identification. Using metabarcoding, Gleason et al (2022) discovered that headwater streams in southern Ontario varied substantially in community composition at very small spatial scales (e.g., over ten meters) and that aquatic macroinvertebrate communities were largely comprised of rare taxa. Gleason et al (2022) detected over 1600 Operational Taxonomic Units (OTUs, a proxy for species) from 149 families, and a single family (Chironomidae) made up one third of the detected OTUs, reflecting the levels of diversity that are not recovered when using family or even genus-level identification.…”

“…Using metabarcoding, Gleason et al (2022) discovered that headwater streams in southern Ontario varied substantially in community composition at very small spatial scales (e.g., over ten meters) and that aquatic macroinvertebrate communities were largely comprised of rare taxa. Gleason et al (2022) detected over 1600 Operational Taxonomic Units (OTUs, a proxy for species) from 149 families, and a single family (Chironomidae) made up one third of the detected OTUs, reflecting the levels of diversity that are not recovered when using family or even genus-level identification. Aquatic macroinvertebrates communities thus contain a high amount of diversity, which is not apparent at coarser taxonomic resolution.…”