Comparison of Host Depletion Protocols and Sequencing Techniques for Urinary Microbiome

Objectives: The goals of this study were to compare the recovered taxa from shotgun metagenomics after two different methods of enriching bacterial compared to host DNA, when compared to an amplification-based technique.

Methods: Voided urine samples were collected via clean catch in clinic (n=10) or through transurethral catheterization in the operating room (n=10). Samples were collected and processed on the same day. Each sample was divided into 3 equal aliquots to allow for parallel processing with three techniques: 1) host cell depletion via centrifuge followed by DNA isolation and WGS; 2) DNA isolation, binding and removal of methylated (host) DNA to enrich microbial DNA (NEBNext); 3) ultrasensitive long-read 16s rRNA gene sequencing through LoopGenomics (Loop-seq). Sterile water +/- lysozyme (negative-controls) and mock microbial communities (positive-controls) were incorporated through all steps of processing. For all WGS data, additional host sequences were bioinformatically removed prior to analysis.

Results: We assessed urine specimens from 20 independent participants (Table). Sequencing data were recovered from all samples processed with the centrifuge technique. When using the NEBNext technique, one catheterized sample failed to provide data. In contrast, the Loop-seq technique failed to provide data for most samples, with only one catheterized and 5 voided samples having taxonomic information. Loop-seq samples that were successfully sequenced had a higher average DNA concentration when compared to samples that did not provide sequencing data (14.127 ng/µL -vs- 2.341 ng/µL, p=< 0.001). Recovered taxa were similar in relative abundance after WGS with both host depletion techniques, and generally consistent with those recovered through Loop-seq, when present. However, Loop-seq showed the potential for contamination (see sample 021) or potential identification bias, as there was a lack of Actinobacteriota, which are known members of the urinary microbiome.

Conclusions: Shotgun metagenomics after host depletion is feasible in urine. The centrifuge technique is more cost-effective but must be performed initially in processing to remove host cells, while the NEBNext host DNA depletion technique may be implemented after DNA isolation. Both methods produced similar taxonomic information, and higher quality data when compared to one method of full length 16S rRNA sequencing that was of limited success.