Be positive: customized reference databases and new, local barcodes balance false taxonomic assignments in...

Area of Science:

• Molecular Ecology
• Bioinformatics
• Genomics

Background:

• Accurate taxonomic assignment in metabarcoding is vital for ecological interpretation.
• Reference database quality and coverage directly impact assignment success.
• Non-target taxa in databases can lead to incorrect species identification.

Purpose of the Study:

• To evaluate the impact of cytochrome c oxidase subunit I (COI) reference database subsets on taxonomic assignment.
• To assess the effectiveness of customized databases for marine metabarcoding data.
• To determine the benefits of incorporating local DNA barcodes into reference libraries.

Main Methods:

• Utilized the COInr database and two subsets (COInr-WO-Insecta, COInr-Med) for COI sequence assignment.
• Applied four different taxonomic assignment algorithms to metabarcoding data from the Mediterranean Sea.
• Quantified changes in unassigned sequences, taxonomic resolution, and assignment accuracy.

Main Results:

• Customized subsets increased unassigned sequences but reduced false assignments.
• Taxonomic resolution improved with more specific, customized databases.
• Adding a small number of local barcodes (0.04% of database) improved resolution for 0.6-1% of ASVs.

Conclusions:

• Customized reference databases are essential for accurate metabarcoding, especially in specific regions.
• Enriching local databases with new barcodes is highly effective for improving taxonomic assignment.
• Balancing database comprehensiveness with regional relevance is key for robust metabarcoding studies.