Bioinformatic pipelines combining denoising and clustering tools allow for more comprehensive prokaryotic and eukaryotic metabarcoding

Environmental DNA metabarcoding is a powerful tool for studying biodiversity. However, bioinformatic approaches need to adjust the diversity of taxonomic compartments targeted as well each barcode gene specificities. We built and tested pipeline based on read correction with DADA2 allowing analysing data from prokaryotic (16S) eukaryotic (18S, COI) life compartments. implemented option cluster amplicon sequence variants (ASVs) into operational units (OTUs) swarm, network-based clustering algorithm, curate ASVs/OTUs using LULU. Finally, assignment was via Ribosomal Database Project Bayesian classifier (RDP) BLAST. validated this ribosomal mitochondrial markers metazoan mock communities 42 deep-sea sediment samples. The results show that ASVs OTUs describe different levels biotic diversity, choice which depends research questions. They underline advantages complementarity LULU-curation producing biodiversity inventories at level approaching one obtained morphological criteria. While removes intraspecific variation, LULU effectively spurious clusters, originating errors or intragenomic variability. Swarm affected alpha beta differently depending genetic marker. Specifically, d-values > 1 appeared be less appropriate 18S metazoans. Similarly, increasing LULU’s minimum ratio proved essential avoid losing species in sample-poor sets. Comparing BLAST RDP underlined accurate assignments can RDP, but highlighted concerted effort build comprehensive, ecosystem-specific databases.