Just keep it simple? Benchmarking the accuracy of taxonomy assignment software in metabarcoding studies.

Area of Science:

• Ecology
• Bioinformatics
• Genomics

Background:

• Metabarcoding enhances biodiversity surveys but relies heavily on accurate taxonomy assignment.
• Existing computational workflows, developed for bacteria, are not always optimal for eukaryotes.
• Accurate species identification is crucial for biodiversity inventories and conservation efforts.

Discussion:

• This study benchmarks various taxonomy assignment strategies for eukaryotic metabarcoding using the COI gene.
• It evaluates methods ranging from sequence similarity to complex probabilistic models.
• The research addresses the critical bioinformatics challenge of naming unknown DNA sequences.

Key Insights:

• Simpler tools (e.g., BLAST, QIIME2 feature classifier) often outperform complex algorithms for higher taxonomic ranks (e.g., family).
• Accurate genus and species-level assignments remain a significant challenge for current algorithms.
• Performance is limited by sparse eukaryotic reference DNA barcode databases.

Outlook:

• The findings highlight best practices for taxonomy assignment in eukaryotic metabarcoding.
• There is a pressing need for community-driven efforts to expand reference DNA barcode databases.
• Future research should focus on improving species-level identification accuracy and database comprehensiveness.