PhyloRef: A Semi-Automated Workflow for eDNA Reference Database Curation via Phylogenetic Anomaly Detection.

Area of Science:

• Bioinformatics
• Genomics
• Ecology

Background:

• Environmental DNA (eDNA) analysis relies on accurate reference databases.
• Public repositories like NCBI often contain annotation errors, misidentifications, and contamination.
• These data quality issues compromise the reliability of biodiversity assessments.

Purpose of the Study:

• To introduce PhyloRef, a semi-automated workflow for curating eDNA reference libraries.
• To enhance the accuracy and reliability of biodiversity assessments using eDNA.
• To address scalability and anomaly detection challenges in reference database curation.

Main Methods:

• PhyloRef employs a phylogeny-guided approach using Snakemake for automation.
• It utilizes taxonomic grouping for scalability and clustering-based anomaly detection.
• The workflow incorporates complete mitochondrial genomes and single-gene sequences, flagging ambiguous records with a "similar_to=" annotation.

Main Results:

• PhyloRef successfully identified and removed anomalous sequences from Chondrichthyes and Actinopterygii datasets.
• Nine chondrichthyan and 401 actinopterygian sequences were removed, improving database quality.
• Ambiguous sequences were flagged, enhancing confidence in downstream eDNA analyses.

Conclusions:

• PhyloRef provides a robust solution for curating eDNA reference databases.
• The workflow significantly improves data quality, leading to more reliable biodiversity assessments.
• Future work includes integrating machine learning and nuclear markers for enhanced resolution.