Oral Microbiome Profiling with mNGS: PaRTI-Seq Workflow
Oral Microbiome and Systemic Health Connections
The oral cavity harbors one of the most diverse microbial ecosystems in the human body — bacteria, fungi, viruses, and archaea coexisting in a dynamic community that influences far more than dental health. Dysbiosis in this microbial landscape has been linked to serious systemic conditions including cardiovascular disease, diabetes, and respiratory infections. The connection between periodontal disease and these downstream health outcomes has driven demand for deeper, more comprehensive oral microbiome profiling.
Traditional approaches such as 16S rRNA gene sequencing and qPCR each carry significant limitations. 16S sequencing is restricted to bacteria and often lacks species-level resolution, missing the viral and fungal components that contribute to dysbiosis. qPCR can detect specific pathogens but only a limited, predetermined panel — leaving unknown or emerging organisms entirely uncharacterized. Metagenomic next-generation sequencing (mNGS) offers the most comprehensive view by capturing all genetic material across all domains of life — but a persistent technical barrier stands in the way.
Oral samples are dominated by human DNA. In standard mNGS workflows, 86–96% of sequencing reads map to the host genome, leaving a thin fraction of sequencing capacity for the microbial community of interest. At these ratios, the majority of sequencing cost is spent reading the human genome rather than characterizing the microbes under investigation — a problem that worsens with saliva and gingival crevicular fluid samples where epithelial cell shedding is high. The PaRTI-Seq workflow addresses this directly through host depletion, dramatically shifting the balance toward microbial reads and enabling deeper characterization of the oral microbiome.
- Periodontal research: mNGS supports species-level bacterial detection beyond broad 16S genus calls.
- Systemic health studies: mNGS gives a broader view of bacterial, viral, and eukaryotic community members.
- Saliva and oral swab workflows: host depletion improves microbial read depth from host-rich specimens.
- Longitudinal intervention studies: preserved individual clustering supports before/after comparisons.
16S vs mNGS for Oral Microbiome Profiling
16S rRNA sequencing remains useful for broad bacterial community surveys, but it is constrained by primer bias, limited species-level discrimination, and bacteria-only detection. Those limits matter when oral microbiome profiling needs to distinguish closely related taxa or detect viruses and eukaryotic organisms that may influence community dynamics.
mNGS removes the marker-gene constraint by sequencing all DNA in the sample. The challenge is efficiency: without host depletion, most reads are spent on human DNA. PaRTI-Seq combines host depletion, ultralow-input library prep, and bioinformatics analysis so oral samples can retain the breadth of mNGS without losing most of the run to host background.
PaRTI-Seq vs Generic mNGS: Microbial Read Enrichment
The core performance advantage of PaRTI-Seq is quantifiable: human DNA contamination drops from 86–96% in generic mNGS to 30–67% with the PaRTI-Seq workflow. This freed sequencing capacity translates directly into microbial signal.
Microbial RPM (Reads Per Million) across three individuals and two swab types:
- Sample A: PU swab increased from 152,068 RPM with generic mNGS to 870,854 RPM with PaRTI-Seq; nylon swab increased from 145,775 RPM to 1,027,592 RPM.
- Sample B: PU swab increased from 371,226 RPM with generic mNGS to 3,296,965 RPM with PaRTI-Seq; nylon swab increased from 626,925 RPM to 2,210,603 RPM.
- Sample C: PU swab increased from 134,118 RPM with generic mNGS to 1,520,185 RPM with PaRTI-Seq; nylon swab increased from 183,724 RPM to 1,847,348 RPM.
PaRTI-Seq increased microbial RPM by 4–9x across all samples and swab types. The highest observed value — 3,296,965 RPM in Sample B — represents nearly nine times the generic workflow result of 371,226 RPM from the same individual. Even the lowest-performing PaRTI-Seq sample (Sample A, PU swab: 870,854 RPM) exceeded the best generic result (Sample B, nylon: 626,925 RPM), demonstrating that the enrichment is consistent enough to guarantee improved depth regardless of starting sample quality.
This enrichment is achieved using the Devin Host Depletion Filter, which selectively removes human genomic DNA prior to sequencing. Because host-depleted samples often yield DNA concentrations below standard quantification limits, the workflow pairs with the Unison Ultralow DNA NGS Library Prep Kit, which maintains reliable library construction even from inputs undetectable by Qubit.
Species Diversity Comparison
Higher microbial read counts translate directly into greater taxonomic resolution. Comparing the best-performing sample (Sample A, PU swab) between workflows:
- Bacteria: PaRTI-Seq detected 1,058 bacterial species, compared with 360 for generic mNGS.
- Viruses: PaRTI-Seq detected 57 viral species, compared with 19 for generic mNGS.
- Eukaryota: PaRTI-Seq detected 11 eukaryotic taxa, compared with 6 for generic mNGS.
- Archaea and fungi: both workflows detected 0 archaea and 0 fungi in this comparison.
PaRTI-Seq detected 1,058 bacterial species compared to 360 with generic mNGS — nearly three times the taxonomic breadth from the same starting sample. Viral detection showed a 2–4x improvement across all samples and swab types, with 57 viral species identified versus 19 with the standard workflow. All species counts represent taxa with 10 or more reads, verified against background controls.
This expanded detection window is particularly relevant for oral microbiome research where low-abundance organisms — including opportunistic pathogens and bacteriophages — may hold clinical significance but remain invisible at shallow sequencing depths.
The 2–4x improvement in viral detection deserves particular attention. Oral viruses, especially bacteriophages, play regulatory roles in microbial community dynamics and can serve as indicators of dysbiosis. At standard sequencing depths without host depletion, most phage diversity falls below the detection threshold. The enrichment provided by PaRTI-Seq brings these organisms into view without requiring additional targeted sequencing panels or specialized bioinformatics.
Swab Type Comparison — PU vs Nylon
A practical consideration for oral microbiome studies is sample collection device selection. The study evaluated two swab types: polyurethane (PU) swabs with a brush head and standard nylon swabs, both placed in 4 ml of 1x PBS immediately after collection.
Both swab types produced comparable results across all metrics — microbial RPM, species diversity, and community composition profiles. Neither swab type introduced systematic bias or showed consistently superior performance. This finding gives researchers flexibility in collection device selection based on availability, cost, or patient comfort without sacrificing data quality.
The consistency between swab types held true for both the PaRTI-Seq and generic workflows, indicating that the host depletion step does not preferentially interact with material from either collection device. For labs standardizing oral sample collection across multiple sites or patient populations, this flexibility simplifies protocol harmonization without introducing a confounding variable.
Workflow Consistency — PCA Analysis
A critical validation for any sample processing workflow is demonstrating that it does not introduce systematic bias into the microbial community profile. PCA analysis of the oral microbiome data confirmed this for PaRTI-Seq.
Principal component analysis (PC1: 41.9%, PC2: 28.6%) revealed clear separation of individual-specific microbial clusters. Samples from the same individual clustered together regardless of whether they were processed through the PaRTI-Seq or generic workflow and regardless of swab type used for collection.
This means the biological signal — the actual microbial community composition unique to each individual — dominates the variance in the data. The host depletion workflow does not distort, flatten, or otherwise alter the community structure. Researchers see the same microbiome profile with dramatically improved depth and resolution.
The practical implication: PaRTI-Seq can be adopted without concern that historical data generated with generic mNGS becomes incomparable. The workflow enhances detection sensitivity while preserving the compositional relationships that define each sample’s microbial signature.
For longitudinal studies tracking oral microbiome changes in response to treatment — periodontal therapy, antibiotic courses, or dietary interventions — this consistency is essential. Researchers can confidently attribute observed shifts in microbial composition to the intervention rather than to processing artifacts introduced by the host depletion step.
