Optimal sequence similarity thresholds for clustering of molecular operational taxonomic units inDNAmetabarcoding studies

Abstract

Clustering approaches are pivotal to handle the many sequence variants obtained in DNA metabarcoding data sets, and therefore they have become a key step of metabarcoding analysis pipelines. Clustering often relies on a sequence similarity threshold to gather sequences into molecular operational taxonomic units (MOTUs), each of which ideally represents a homogeneous taxonomic entity (e.g., a species or a genus).However, the choice of the clustering threshold is rarely justified, and its impact on MOTU over-splitting or over-merging even less tested. Here, we evaluated clustering threshold values for several metabarcoding markers under different criteria: limitation of MOTU over-merging, limitation of MOTU over-splitting, and trade-off between over-merging and over-splitting. We extracted sequences from a public database for nine markers, ranging from generalist markers targeting Bacteria or Eukaryota, to more specific markers targeting a class or a subclass (e.g., Insecta, Oligochaeta).Based on the distributions of pairwise sequence similarities within species and within genera, and on the rates of over-splitting and over-merging across different clustering thresholds, we were able to propose threshold values minimizing the risk of oversplitting, that of over-merging, or offering a trade-off between the two risks. For generalist markers, high similarity thresholds (0.96-0.99) are generally appropriate, while more specific markers require lower values (0.85-0.96). These results do not support the use of a fixed clustering threshold. Instead, we advocate careful examination of the most appropriate threshold based on the research objectives, the potential costs of over-splitting and over-merging, and the features of the studied markers.

Evaluating the sampling effort for the metabarcoding‐based detection of fish environmental DNA in the open ocean

“…Owing to this algorithm, QCauto does not require a minimum similarity threshold to distinguish taxa (Tanabe & Toju, 2013). As a result, QCauto may be robust against splitting and lumping errors, which are unavoidable in ordinal similarity‐based clustering or discrimination methods with a single threshold value (Bonin et al, 2022). Furthermore, among the various assignment algorithms, QCauto returns the most reliable results when the completeness of the reference sequence database of all potentially observable species is low (Tanabe & Toju, 2013), as expected in the open ocean.…”

Metabarcoding data reveal vertical multitaxa variation in topsoil communities during the colonization of deglaciated forelands

“…Fifth, we clustered sequences at a threshold of 96% (Bact02, Euka02, Inse01), 95% (Fung02), 92% (Olig01) or 85% (Coll01) sequence similarity using the sumaclust program (https://git.metabarcoding.org/obitools/sumaclust/wikis/home). These thresholds were selected for each taxon on the basis of the distributions of pairwise sequence similarities within and between species of the same genus for each marker (Bonin et al, 2022). The selected thresholds allow us to minimize the risk that multiple individuals belonging to the same species are assigned to different molecular operational taxonomic units (MOTUs), while limiting the probability that distinct species are collapsed in one single MOTU (Bonin et al, 2022).…”

“…These thresholds were selected for each taxon on the basis of the distributions of pairwise sequence similarities within and between species of the same genus for each marker (Bonin et al, 2022). The selected thresholds allow us to minimize the risk that multiple individuals belonging to the same species are assigned to different molecular operational taxonomic units (MOTUs), while limiting the probability that distinct species are collapsed in one single MOTU (Bonin et al, 2022). Differences in thresholds are related to differences in taxonomic resolution across markers.…”

“…The selected thresholds allow us to minimize the risk that multiple individuals belonging to the same species are assigned to different molecular operational taxonomic units (MOTUs), while limiting the probability that distinct species are collapsed in one single MOTU (Bonin et al, 2022). Differences in thresholds are related to differences in taxonomic resolution across markers.…”

Towards exhaustive community ecology via DNA metabarcoding

“…For instance, if freshwater biodiversity is analysed using primers amplifying bacteria, diatoms, molluscs, insects, fishes and amphibians, key taxa such as crustaceans and most microeukaryotes will remain undetected. Furthermore, integrating the results of multiple markers to obtain a coherent, homogeneous species lists can be challenging (Bonin et al, 2023; Jurburg et al, 2021; see Section 3.3).…”