Why context matters in microbiome science
A microbiome result means little in isolation. Reference cohorts give each measurement a population to be read against — and define the boundaries of what that comparison can support.
There is no single "normal" microbiome, so every VAMS BIOME result is contextualized against reference cohorts rather than a fixed ideal. This page explains how cohorts are built, how yours is matched, and the statistical framework behind percentile context. It also states the boundaries plainly: cohort comparison describes where your sample sits relative to a population — an association, not a diagnosis, and not a prediction about your health.
Why cohorts matter
Comparison is what turns a raw measurement into meaning — and choosing the right comparison group is itself a scientific decision.
The same abundance value can read very differently depending on who it is compared against. Reference cohorts provide that comparison group, letting a result be expressed as percentile context rather than an absolute judgment. We are transparent about cohort composition and its limits: population diversity is incomplete, cohorts evolve, and a difference from a reference is an observation to explore, not evidence of a condition. This restraint is what makes cohort context trustworthy.
Derived from sequencing data. For educational and informational use only; not intended for diagnosis or treatment.
Population Diversity
Population Diversity explains the scientific role this topic plays within VAMS BIOME’s microbiome intelligence architecture.
Population Diversity explains why microbiome observations become more meaningful when they are interpreted against ecosystem-specific reference distributions rather than viewed as isolated findings.
Cohort Matching
Cohort Matching explains the scientific role this topic plays within VAMS BIOME’s microbiome intelligence architecture.
Cohort Matching explains why microbiome observations become more meaningful when they are interpreted against ecosystem-specific reference distributions rather than viewed as isolated findings.
Statistical Framework
Statistical Framework explains the scientific role this topic plays within VAMS BIOME’s microbiome intelligence architecture.
Statistical Framework explains why microbiome observations become more meaningful when they are interpreted against ecosystem-specific reference distributions rather than viewed as isolated findings.
Interpretation Boundaries
What a comparison can — and can't — mean.
Comparing your sample to a reference population tells you how common your pattern is — for example, that a feature sits around the 70th percentile means it is higher than most people in that reference group. It does not mean you are healthier or unhealthier, and it is not a risk score for any condition. Reference data also has limits: cohorts may not perfectly match your age, geography or life stage, so comparisons are best read as context, not verdicts.
- ✓Population context
- ✓Ecosystem-specific distributions
- ✓Prevalence benchmarks
- ✓Percentile interpretation
- ✓Confidence-aware reporting
Cohort Visualization
A visual section designed for fast scanning, comparison, and guided exploration.
Cohort Visualization explains why microbiome observations become more meaningful when they are interpreted against ecosystem-specific reference distributions rather than viewed as isolated findings.
Understand how your results are contextualized
See the cohorts, the matching logic, and the statistical boundaries that shape every percentile — so you know exactly what a comparison does and does not claim.
- ✓ Population context
- ✓ Ecosystem-specific distributions
- ✓ Prevalence benchmarks
- ✓ Percentile interpretation
- ✓ Confidence-aware reporting
Derived from sequencing data. For educational and informational use only; not intended for diagnosis or treatment.
Method validation
Curation & quality
How reference datasets are assessed before admission.
| Curation scoring | Each dataset is scored for metadata completeness, sample size, geography fit, sequencing compatibility and confounder risk. |
|---|---|
| Admissibility | Datasets meeting curation thresholds are admitted to the reference set; curated datasets carry a mean admissibility score of 4.4 out of 5. Registry-derived |
| Harmonisation | Cohorts are harmonised across body sites, conditions and sequencing approaches under standardised definitions. |
Reference set
Reference cohort coverage
Derived from the curated Phase 1 cohort & dataset registry. Figures are counts from the registry, not marketing estimates.
| Curated datasets | 106 datasets across 58 curated cohorts |
|---|---|
| Expected reference samples | Approximately 23,000 Sum of dataset expected sample counts |
| Body-site systems | 5 — gut, vaginal, oral, skin, and multi-system |
| Datasets by body site | Gut 47 · vaginal 22 · skin 19 · oral 11 · multi-system 7 |
| Condition categories | 35 curated cohort definitions across the four ecosystems |
| Healthy reference cohorts | 13 — global and regional healthy baselines |
| Sequencing approaches | 16S amplicon (79 datasets) and shotgun metagenomics (21 datasets) |