Impact of DNA isolation method on a microbial community standard assessed by metagenomics nanopore sequencing: implications for vaginome research

Background/Objectives: The impact of an unhealthy vaginal microbiome (vaginome) is enormous and increases the risk of sexually transmitted infections, such as human papillomavirus (HPV). Shotgun metagenomic sequencing provides advantages over amplicon sequencing for analyzing the vaginome, but is a challenge to perform in lower microbial biomass samples with high human DNA content such as cervical scrapes. The ZymoBIOMICS microbial community standard is a well-defined mock community consisting of 3 Gram-negative bacteria (36%), 5 Gram-positive bacteria (60%) and 2 yeast strains (4%) with varying cell wall composition enabling validation of DNA extraction methods. In this study, the impact of three DNA extraction methods and 2 bioinformatics pipelines was
evaluated on the community standard and cervical scrapes stored in PreservCyt buffer (Hologic).
Methods: After DNA extraction with the automated eMAG (Biomérieux), SwiftX DNA and SwiftX Hi-Sense (Xpedite Diagnostics) samples have been sequenced using the Native barcoding kit V14, flow cell R10.4.1 and MinION Mk1B. DNA yield was measured with Qubit dsDNA high sensitivity assay kit. Data was analyzed with both the EPI2ME Labs Metagenomic workflow (Oxford Nanopore Technologies) and the BugSeq Metagenomic sequencing pipeline.
Results: For the microbial community standard (30µl), DNA yield was 269ng with the eMAG method versus 127ng and 47ng for SwiftX DNA and SwiftX Hi-Sense, respectively. However, for the cervical scrape (3mL) eMAG resulted in a DNA concentration of 824ng and both Swift methods resulted in over 2300ng. Although the DNA yield for eMAG method was lower, this extraction method resulted in at least 4 times more nanopore reads for the clinical sample. Approximately, 94% of the obtained nanopore reads from the cervical scrape isolated with the automated eMAG system were of human origin. DNA extraction of the microbial community standard with eMAG, SwiftX DNA and SwiftX Hi-Sense resulted in 86%, 49% and 35% Gram negative bacterial reads, respectively.
Furthermore, both Swift methods resulted in 36% (SwiftX) and 55% (Swift Hi-Sense) reads of other microorganisms (as compared to what is in the standard) or unclassified reads with the EPI2ME workflow. The BugSeq workflow showed similar results for the eMAG sample, however for the both Swift extractions a maximum of 18% reads of other microorganisms or unclassified reads was found.
More Gram positives (35%) were recovered with BugSeq.
Conclusions: DNA extraction method and bioinformatics analysis have an impact on sequencing yield and species identification, indicating that both lab-based and bioinformatics methods need optimization. All extraction methods used show bias for Gram-negative
bacteria. Currently, experiments are being performed with the ZymoBIOMICS DNA Miniprep kit as this lysis approach claims to eliminate bias associated with unequal lysis efficiencies of different micro-organisms. Fortunately, the vaginome does not contain a large diversity of microorganisms and is predominantly composed of Gram-positive bacteria, the bias towards Gram-negative bacteria may not play a significant role.