HAYSTAC: A Bayesian framework for robust and rapid species identification in high-throughput sequencing data.

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Accurate species identification in metagenomic samples is crucial for various applications.
• Existing tools often demand significant computational resources and can produce false positives.

Purpose of the Study:

• To introduce High-AccuracY and Scalable Taxonomic Assignment of MetagenomiC data (HAYSTAC), a novel tool for metagenomic data analysis.
• To develop a user-friendly and computationally efficient method for taxonomic assignment.

Main Methods:

• HAYSTAC utilizes a Bayesian framework for inferring species abundance and statistical support.
• It employs competitive read mapping with user-constructed databases from public genomes.
• The tool is designed to handle ancient and modern DNA, as well as incomplete reference databases.

Main Results:

• HAYSTAC demonstrated fewer false positives compared to Kraken2/Bracken, KrakenUniq, and MALT across all simulations.
• It showed fewer false positives than Sigma in ancient DNA simulations.
• HAYSTAC requires less memory for both database construction and sample analysis than competing methods.

Conclusions:

• HAYSTAC provides a highly accurate and efficient solution for taxonomic assignment in metagenomic data.
• Its ability to handle diverse data types and incomplete databases makes it versatile for hypothesis-driven research.
• The tool is suitable for analyzing both simulated and empirical ancient metagenomic datasets.