Host DNA removal via chemical depletion for improved shotgun metagenomic analysis

Host DNA can present challenges for microbiome studies utilizing whole genome sequencing (WGS). Notably, host DNA can influence the number of bacterial reads that can be classified during downstream analyses. This is a common issue for samples collected from sites with low bacterial burden such as the upper nasal passages, mouth, skin, or vagina. The ability to identify microbial communities with accuracy and sensitivity from such environments could be significantly improved with a robust method for host DNA removal. Propidium monoazide (PMA), a photo-reactive DNA-binding dye used for the detection of live and or dead cells in a sample, has shown promising results for host DNA depletion. In a previous study, selective osmotic lysis and PMA was used to remove mammalian DNA from human samples. PMA was found to perform more efficiently than other commercial host depletion kits. To evaluate the effectiveness of selective osmotic lysis and PMA treatment on clinical samples, we performed WGS on a saliva sample collected from one participant. This sample was divided into two treatment groups: No PMA and PMA. To test if storage conditions might also influence sequence read coverage, each treatment group was split into two subgroups: frozen and frozen with glycerol. We hypothesized that host DNA removal via selective osmotic lysis and PMA treatment would significantly reduce human DNA reads, improving results. We found that this novel host depletion method removed 98% of identified human reads across storage conditions, remarkably improving bacterial relative abundance at the species-level