StrainMash

What does it do?

Use StrainMash to identify the RefSeq type strain most similar to a draft genome assembly.

StrainMash uses the MinHash algorithm implemented by Mash to compare an assembly quickly with reference type strains. It is useful when the expected species of an assembled isolate is uncertain.

StrainMash identifies the closest reference in its database; it does not by itself provide the multi-method rMLST and ANI evidence produced by Unknown Isolate, and it is not a mixed-community profiler.

How do I use it?

Subject

In the Subject field, enter:

StrainMash

Spelling matters, but matching is not case-sensitive.

Description

In the Description field, enter one assembly SEQID per line:

2026-SEQ-0001
2026-SEQ-0002

Attachments

No attachment is required. StrainMash retrieves the draft assembly associated with each requested SEQID.

Optional parameters

The supplied documentation does not identify optional parameters for the Redmine StrainMash automator.

Example

See issue 12860 for an example StrainMash request.

Interpreting results

When StrainMash finishes, it uploads a ZIP archive containing an output directory. The directory contains one text report per requested SEQID.

Each report contains six columns.

MashDistance

The existing automator documentation describes this value as a closeness score for the query and reference and recommends at least 0.98 for a very close match.

Standard Mash distance convention normally uses smaller values for closer sequences, so the reported field may be a transformed similarity value or may be mislabeled in the legacy documentation. Confirm the current output semantics before using the 0.98 threshold in a final identification decision.

NumMatchingHashes

Reports the number of matching hashes. The documented range uses 1000/1000 as a perfect match and 0 as no similarity. The legacy guidance recommends more than 850 matching hashes for a close reference match.

MedianMultiplicity

The legacy documentation states that this value can usually be ignored and should normally be 1.

Pvalue

Reports the statistical significance assigned to the hit. The legacy guidance expects 0 for a very close reference match.

ReferenceStrain

Reports the NCBI accession of the reference strain.

Organism

Reports the organism name associated with the reference strain.

Interpret the closest match using the score semantics, matching-hash support, reference coverage, and organism context. A closest database match is not necessarily a definitive species identification when reference representation is limited or the assembly is contaminated, incomplete, or divergent.

How long does it take?

StrainMash generally takes approximately 30 seconds to one minute per sample. Total runtime depends on sample count and service workload.

What can go wrong?

A requested SEQID is unavailable

Symptom: The Redmine issue receives a warning identifying unavailable sequences.

Likely cause: StrainMash cannot locate a draft assembly for the requested SEQID.

What to do: Verify each SEQID, confirm that its assembly is available, and submit a corrected request.

The closest match has weak support

Symptom: The top reference result has relatively few matching hashes or an uncertain similarity score.

Likely cause: The query may be distant from represented type strains, poorly assembled, contaminated, or absent from the reference collection.

What to do: Review assembly quality and contamination, then use Unknown Isolate for broader rMLST, MASH, ANIb, and ANIm evidence.

MashDistance appears inconsistent with Mash conventions

Symptom: The reported value increases with similarity even though standard Mash distance normally decreases as sequences become more similar.

Likely cause: The automator may report a transformed similarity score under the MashDistance heading, or the legacy description may be incorrect.

What to do: Confirm the current output calculation or inspect a known reference match before interpreting this field quantitatively.

  • Unknown Isolate — combines rMLST, MASH, ANIb, and ANIm evidence for uncertain isolate identification.
  • AutoCLARK — reports species represented in raw reads or draft assemblies.
  • Kraken2/Bracken and MetaPhlAn4 — profile taxonomic composition in metagenomic data.