Depth masking is not a technical detail
Why low-depth positions, primer dropouts, and ambiguous bases should be handled visibly when producing consensus genomes for surveillance and phylogeny.
A consensus genome is an interpretation
A consensus FASTA looks simple, but it carries many decisions. Which reads were used? Were primers trimmed? What depth was required for a base? Were low-quality variants filtered? Were low-coverage positions masked as N? These choices shape downstream phylogeny, lineage calls, and mutation tables.
If low-depth sites are not masked, a consensus can look cleaner than the evidence supports. That can create false precision in trees and misleading mutation summaries. In outbreak work, where people may compare samples by a small number of differences, this matters.
Masking protects interpretation
Masking is sometimes treated like a technical nuisance, but it is actually a statement of honesty. It tells the reader that the workflow refuses to call a base without enough evidence. The exact threshold depends on pathogen, platform, amplicon design, expected diversity, and reporting use case.
For short-read viral pipelines, I often start with a minimum depth threshold and then inspect breadth, ambiguous base percentage, and dropout patterns. For amplicon workflows, primer failure can create structured missingness, so a simple average depth is not enough.
What I would publish with each consensus
- Median depth and breadth of coverage at the selected depth threshold.
- Percent Ns or ambiguous bases in the final consensus.
- Primer scheme, trimming method, and reference accession.
- Variant-calling filters and masking threshold.
- QC status: pass, review, fail, or limited-use sequence.
Why this improves communication
When a report states the masking logic, the downstream discussion becomes more scientific. A weak sequence can still be useful for partial evidence, but it should not be overinterpreted as a complete genome. A strong sequence can be used more confidently in lineage assignment and phylogenetic context.
This is the kind of small methods discipline that improves a whole surveillance system. It makes results easier to compare across batches, laboratories, and time.
Primer trimming and masking are connected
Depth masking is often discussed after consensus generation, but the decision begins earlier. Primer trimming, read filtering, duplicate handling, mapping quality, and variant-calling thresholds all affect which positions appear supported. If primers are not handled correctly, false confidence can appear near amplicon edges.
This is one reason amplicon workflows need a primer scheme file, clear trimming steps, and a depth report. When a consensus is used for phylogeny or mutation interpretation, the analyst should be able to explain which positions were called, masked, or ignored.
How I classify sequence usability
Not every consensus has the same use. A high-coverage, low-ambiguity genome can support lineage assignment, tree placement, and mutation summaries. A partial genome may support target confirmation or broad genotype context but not fine-scale transmission inference. A low-depth sequence may belong only in an internal follow-up note.
That classification should be visible. I prefer labels such as pass, review, limited-use, and fail because they communicate confidence without burying the reader in raw tables. The label should always be traceable to depth, breadth, ambiguity, and control performance.
- Pass: high breadth, adequate depth, clean controls, low ambiguity.
- Review: useful sequence with one or more caveats that need manual inspection.
- Limited-use: partial evidence suitable only for restricted interpretation.
- Fail: insufficient support for genomic interpretation.