ParaRef: a decontaminated reference database for parasite detection in ancient and modern metagenomic datasets.
Jonas Niemann, Yuejiao Huang, Liam T Lanigan, Arve L Willingham Grijalba, Robert R Dunn, Martin Sikora, Hannes Schroeder
Abstract
Open AccessShotgun metagenomics holds great potential for identifying parasite DNA in biological samples, but its effectiveness is limited by widespread contamination in publicly available reference genomes, which hinders accurate detection. In this study, we systematically quantify and remove contamination from 831 published endoparasite genomes to create ParaRef, a curated reference database for species-level parasite detection. We show that decontamination significantly reduces false detection rates and improves overall detection accuracy. Our study highlights the pervasive issue of contamination in public databases and offers a resource that will enhance the reliability of parasite detection using metagenomics.