We conducted six metabarcoding studies from different aquatic habitats. It has been confirmed that fungal metabarcoding has the advantages of overcoming the cultivation bias, increasing the number of possible samples by high-throughput sequencing, and enabling the detection of yet uncultivated fungal taxa. However, the already well-known limitations, in particular the lack of taxonomic accuracy when using automated classification also occurred. During the manual curation of pipeline-retrieved datasets, the quality of the automatic classification, specifically phylogenetic rank and current nomenclature, was verified. The manual curation step revealed that only 10–20% of the Operational Taxonomic Units (OTUs) were identified correctly by the automatic classification. Re-blasting of the remaining OTUs resulted in taxonomic and nomenclatural correction of a further 30–40% of OTUs. Approximately 50% of the OTUs resulted in insufficient identifications, which can be attributed to shortcomings in the database, including yet unidentified species, missing reference sequences, and the disadvantages of the genetic marker used. Pipeline parameters that can be adjusted to improve the classification of taxa in metabarcoding studies include adjusting the cut-off threshold for sequence similarity, updating and refining databases, and aligning and collapsing of OTUs to species hypotheses. Here, we describe a method of manual data curation and its impact on improving the taxonomic resolution of a metabarcoding dataset. Shortcomings as well as advantages are discussed to illustrate the need for bioinformatic tools that support the process of data curation. • Metabarcoding studies of fungal communities often lack sufficient taxonomic accuracy. • Species hypotheses and current nomenclature can be curated in the retrieved datasets. • Curation improves the identification of communities at the genus and species level. • A protocol for curating metabarcoding datasets from fungal communities is provided.
Christiane et al. (Sun,) studied this question.