Team Presents Strategy for Speedy Species Detection in Metagenomic Sequence Data

For a paper in PLOS Computational Biology, investigators at the University of Oxford, Ludwig Maximilian University, the University of Copenhagen, and other centers present a computational strategy for quickly ascribing high-throughput metagenomic sequences from ancient or contemporary samples to a specific species with reduced computational memory needs — an approach dubbed "high-accuracy and scalable taxonomic assignment of metagenomic data" (HAYSTAC). After showing that the Bayesian framework-based method compared favorably to existing approaches for identifying species in simulated short-read sequence data, leading to fewer false-positive identifications, the team used HAYSTAC to analyze two available metagenomic datasets produced from ancient human bone or ancient dental calculus samples. "[W]e present a tool that can robustly assess whether a specific species is present in a metagenomic sample in one step, without the need to combine different pipelines to validate the results," the authors note, adding that HAYSTAC "reliably produces the lowest number of false positive identifications, making it a valuable tool for both ancient and modern DNA microbial research."