In response to

"Should I use V3-V4 or V4? What's the difference?"

These refer to specific stretches of the 16S ribosomal RNA gene, and the stretch you amplify shapes what organisms you detect, how accurately you can classify them, and which public datasets your results are comparable to.

The 16S Gene

The 16S rRNA gene (~1,542 bp in most bacteria) encodes the small subunit of the prokaryotic ribosome. It is present in essentially all bacteria and archaea, which is why it became the universal marker gene for culture-independent community profiling. The gene has a consistent alternating structure: conserved regions flanking variable regions. The conserved regions are where universal primers bind; the variable regions (nine of them, V1 through V9) accumulate mutations at rates that make them phylogenetically useful.

The key insight is that you are never sequencing the whole gene with short-read Illumina; you are always amplifying one or two of these variable regions using primers anchored in the flanking conserved regions. Which region(s) you capture is the single most consequential methodological choice in an amplicon study.

The 16S rRNA gene with variable regions V1–V9 marked at their approximate positions (E. coli K-12 numbering). Click any region for details. Conserved flanking sequences (gray) are where universal primers bind.

Each variable region differs in length, nucleotide variability, and how much phylogenetic information it carries at different taxonomic ranks. V1 and V9 are both short (~31 bp) and high-variability but are rarely useful alone. The intermediate regions (V3, V4, and V5 in particular) sit in the practical sweet spot for short-read amplicon work.

The positions above use the standard E. coli K-12 reference coordinates. The actual lengths of these regions vary across lineages, which is part of what makes them variable regions. The numbers are a convention, not an absolute.

Common Options

Taxonomic Resolution

The heatmap below shows approximate classification success rates at each taxonomic rank for common region choices, based on published benchmarking studies. All short-read regions perform well at phylum level; the differences become meaningful at family and genus.

Approximate classification rates against SILVA-trained classifiers, synthesized from benchmarking literature.1,4,5 Values reflect typical performance across mixed-community mock standards and field samples; actual rates vary by sample type, database version, and classifier. V4 rates may be slightly underestimated here: in practice, the depth of V4 classifier training data often partially compensates for its lower information content relative to V3–V4.

Primer Bias

The primers matter as much as the region. The original 515F/806R pair used in the Earth Microbiome Project underestimates SAR11 (one of the most abundant marine bacteria on Earth) and some Thaumarchaea due to single-nucleotide mismatches in those lineages.3 Two modifications address this: the 806Rb (Apprill et al., 20156) revision corrects the SAR11 mismatch, and the 515F-Y (Parada et al., 20153) modification improves Thaumarchaea coverage. The EMP now recommends using both 515F-Y and 806Rb together.

No primer pair achieves truly universal coverage. The V3–V4 pair (341F/805R) has slightly lower archaea coverage than V4-targeted primers, which matters for environments like hot springs, deep subsurface, and some soils where Archaea are ecologically important. Klindworth et al.1 is the most comprehensive in silico evaluation of bacterial and archaeal primer coverage to date and is worth reading if you are designing a new study in an unusual system.

Plastid and mitochondrial contamination is another form of bias worth noting: host plant 16S (chloroplast 16S is homologous to bacterial 16S) and mitochondrial 12S can co-amplify with some primer pairs. For root microbiome, phyllosphere, or insect gut studies, the 799F forward primer is specifically designed to reduce plastid amplification.

Platform Considerations

On Illumina, you are constrained by read length. The tradeoffs are straightforward:

Target Amplicon Min. read config
V4 (515F/806Rb) ~253 bp 2×150 (comfortable overlap)
V4–V5 (515F-Y/926R) ~411 bp 2×250 (minimal overlap)
V3–V4 (341F/805R) ~464 bp 2×300 (required)
Full-length ~1,465 bp Not feasible

Paired-end reads from both ends of the amplicon must overlap in the center to be merged into a single sequence. That overlap region is sequenced from both directions, which is what enables error correction in tools like DADA2. A minimum of ~20 bp is required; ~50 bp is more reliable in practice. As amplicons get longer, reaching adequate overlap requires longer reads, and at some point the combination stops working entirely.

Each row shows the amplicon span (bracket) with R1 (blue, extending from the forward primer) and R2 (orange, extending from the reverse primer) at proportional length. Green = overlap region sequenced from both directions; red dashes = unsequenced gap that prevents merging. Overlap = 2 × read length − amplicon length.

On PacBio (CCS / HiFi) and Oxford Nanopore, full-length 16S is realistic. Callahan et al.5 showed that PacBio CCS with DADA2 can sequence the full ~1,500 bp gene at near-zero error rates, achieving single-nucleotide resolution and approaching species-level discrimination. Nanopore accuracy has improved substantially but still trails PacBio CCS for this application as of 2025.

Approximate genus- and species-level classification accuracy synthesized from benchmarking studies (Callahan et al. 2019;5 Johnson et al. 20194). Short-read approaches plateau near genus level; only full-length sequencing approaches species-level discrimination. Nanopore accuracy continues to improve with successive chemistry and basecalling releases.

Which Region Should I Use?

References

  1. Klindworth A, Pruesse E, Schweer T, et al. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Research. 2013;41(1):e1. doi:10.1093/nar/gks808
  2. Caporaso JG, Lauber CL, Walters WA, et al. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. The ISME Journal. 2012;6(8):1621–1624. doi:10.1038/ismej.2012.8
  3. Parada AE, Needham DM, Fuhrman JA. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environmental Microbiology. 2016;18(5):1403–1414. doi:10.1111/1462-2920.13023
  4. Johnson JS, Spakowicz DJ, Hong B-Y, et al. Evaluation of 16S rRNA gene sequencing for species and strain-level microbiome analysis. Nature Communications. 2019;10:5029. doi:10.1038/s41467-019-13036-1
  5. Callahan BJ, Wong J, Heiner C, et al. High-throughput amplicon sequencing of the full-length 16S rRNA gene with single-nucleotide resolution. Nucleic Acids Research. 2019;47(18):e103. doi:10.1093/nar/gkz569
  6. Apprill A, McNally S, Parsons R, Weber L. Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquatic Microbial Ecology. 2015;75:129–137. doi:10.3354/ame01753