Designing a microbiome study: depth, primer choice, and where expert advice pays off

Two design decisions set the ceiling on what your microbiome study can detect: how deep you sequence each sample, and which 16S region you target. Get them wrong and no clever analysis afterwards can recover what was missed. Here is how to think about both, and where the value of an experienced bioinformatics partner shows up.

Two decisions that lock everything else

A microbiome study has hundreds of decisions, and most are reversible. Software versions, statistical tests, even the visualisation layout, all can be revisited after the data lands. Two cannot. How deep you sequence each sample, and which 16S region you target.

Both happen before any bioinformatics begins, and both set a hard ceiling on what the analysis can detect afterwards. Get either wrong and no downstream pipeline will recover what the experiment never captured. The two decisions are also coupled, the right depth depends on the region, and vice versa.

How deep is enough? Sequencing depth

The right depth depends on three things: how complex the sample is, how fine the taxonomic resolution needs to be, and what comparison you intend to make. A few empirical anchors:

Human gut samples are relatively low-complexity. 10,000–50,000 reads per sample is usually enough for genus-level diversity; 50,000–100,000 if you want species-level resolution via denoising algorithms. The American Gut Project found genus-level rarefaction curves plateaued around 25,000 reads per sample (McDonald et al., 2018).

Soil and marine samples are far more diverse. Recovering rare taxa and supporting robust beta-diversity comparisons typically requires 100,000–500,000 reads per sample. Lundberg et al. (2012) showed roughly 200,000 reads were needed to identify rare taxa in rhizosphere soil.

Fungal ITS regions vary in length and copy number across taxa, so under-sequencing distorts the picture, plan for 30,000–100,000 reads per sample. Other markers like COI or rbcL push higher: 50,000–150,000 for environmental work, given the variability and lack of universal primers.

The signal that you have sequenced deep enough is empirical: the rarefaction curve should plateau before the depth you have chosen. Over-sequencing wastes budget; under-sequencing leaves rare-taxa questions unanswerable.

Which region? 16S primer choice

The 16S ribosomal RNA gene contains nine hypervariable regions (V1–V9), separated by conserved segments. Different regions resolve different organisms at different taxonomic levels, the choice shapes what your assay can see.

V3–V4 is the most-used region in microbiome and environmental work, balanced length, strong species-level discrimination across many taxa. V4 alone is shorter, fits efficiently into high-throughput Illumina runs, and behaves robustly across diverse taxa, a defensible default for large cohorts. V1–V3 gives slightly better species-level discrimination in gut microbiome work specifically. V5–V6 helps environmental samples where Firmicutes and Actinobacteria need finer resolution.

For fungi, ITS replaces 16S; for eukaryotic profiling, 18S. And a modern analysis chooses Amplicon Sequence Variants (ASVs) over the older OTU clustering approach, for better resolution and reproducibility.

How depth and region interact

The two decisions are coupled. A short region like V4 produces tightly defined ASVs that need less depth to characterise; a longer region like V3–V4 produces more variants and benefits from deeper coverage to keep rare ASVs above the noise floor. ITS amplicons vary in length, so depth must compensate for uneven amplification across taxa.

The right combination is rarely formulaic. It depends on sample type, the statistical question being asked, the cohort size, and what the data needs to support downstream, publication, an operational dashboard, or a regulatory submission.

Where Nova Lifetech changes the outcome

These are not look-up-the-answer decisions. They depend on the cohort, the statistics, the budget, and the eventual destination of the data. We have designed and run hundreds of microbiome studies across all four sample classes, gut, soil, water, fungal. That experience compresses into specific guidance for your project: which region for your sample type, what depth to commit to, where to pilot, and which trade-offs are safe to accept.

We work alongside the principal investigator or the operations lead from study design through data delivery. The handoff at the end of the engagement is not just data. It is the rationale for every decision, so your team can defend it in review, adjust it for the next round, or scale it without consulting us again.

Pilot first

The cheapest insurance against the wrong depth or wrong region is a small pilot, five to ten samples at multiple depths, or with two candidate regions, before committing the full cohort. The pilot answers the depth question empirically (rarefaction stabilises or it does not) and validates that the chosen region actually distinguishes the taxa your study cares about.

A scoped pilot costs less than re-running a single missed cohort. Most clients only need one.

↗Sequencing depth: practical anchors for tropical samples
↗Amplicon and metagenomics services
↗Talk to us
Two decisions that lock everything else
How deep is enough? Sequencing depth
Which region? 16S primer choice
How depth and region interact
Where Nova Lifetech changes the outcome
Pilot first

AMR surveillance dashboard: a live project

An interactive dashboard for a national AMR repository: resistance samples across regions, organisms and drug classes, and a live example of what we deliver alongside a typical engagement.

Eukaryotic genome annotation: a sample report

A snail-plot quality report from a eukaryotic genome assembly engagement, paired with completeness, scaffold statistics, and composition. The shape we hand off alongside the methods writeup.

Tell us what you are working on.

A short note about your samples and questions is enough to start. We reply with how we would approach it.