Decoding diversity: evaluation of metabarcoding pipelines using a Sanger-verified Cladonia natural community as reference

Metabarcoding of environmental DNA is widely used to assess fungal diversity, yet the choice of bioinformatic pipeline strongly influences results. Here we compare five pipelines to detect Cladonia diversity from soil eDNA across Europe, using an extensive floristic survey combined with Sanger sequencing as a reference. All pipelines similarly detected dominant species but differed in their recovery of rare taxa, hidden diversity, or putative artefacts. The total number of operational taxonomic units (OTUs) varied 2.5-fold across pipelines, and overlap with the reference dataset ranged from 21% to 40%. The pipeline by Bálint et al. (2014), followed by Swarm v2 clustering and VSEARCH chimera filtering, achieved the greatest agreement with the taxonomic survey and recovered the highest number of rare OTUs. The DADA2-based pipeline with DECIPHER clustering yielded the highest number of OTUs not found by the taxonomic survey. In contrast, LULU curation yielded the most conservative dataset, with the lowest diversity and lowest overlap with the Sanger-verified reference.