dRep
What does it do?
Use dRep to compare genome assemblies by average nucleotide identity (ANI) or to dereplicate a genome set.
dRep first performs Mash-based primary clustering. It then applies a selected secondary comparison algorithm only within genome clusters meeting the primary Mash ANI threshold. This two-stage design avoids expensive ANI comparisons between clearly unrelated genomes.
The compare command produces relatedness and ANI outputs. The dereplicate command groups similar genomes and selects representatives according to dRep's workflow.
For background, see the dRep documentation and Olm et al. (2017). Cite the dRep authors when publishing results produced with this automator.
How do I use it?
Subject
In the Subject field, enter:
drep
Spelling matters, but matching is not case-sensitive.
Description
The first line selects the analysis mode:
analysis=custom
Enter optional parameters next, followed by one assembly SEQID per line:
analysis=custom
2026-SEQ-0001
2026-SEQ-0002
2026-SEQ-0003
Only custom is documented as currently supported. It compares or dereplicates only the listed genomes.
Attachments
The supplied documentation does not identify attachment support. Use available assembly SEQIDs unless the current implementation has been verified to accept files.
Optional parameters
mash_ANI_threshold
Sets the primary Mash ANI clustering threshold. Only genomes clustering at or above this threshold proceed to secondary ANI comparison.
- Default:
0.9(90%) - Example:
mash_ANI_threshold=0.8
Lowering this threshold can create broader clusters and substantially increase secondary-comparison runtime.
S_ANI
Sets the secondary clustering ANI threshold.
- Default:
0.99(99%) - Example:
S_ANI=0.9
coverage_threshold
Sets the minimum overlap required between genomes during secondary comparisons.
- Default:
0.1(10%) - Example:
coverage_threshold=0.2
clusteralgorithm
Selects the hierarchical clustering linkage method.
- Default:
average - Example:
clusteralgorithm=ward - Supported values:
medianweightedsinglecompleteaveragewardcentroid
comparisonalgorithm
Selects the secondary ANI comparison algorithm.
- Default:
ANImf - Example:
comparisonalgorithm=fastANI - Supported values:
fastANIANImfANIngANIgoANI
command
Selects comparison or dereplication behavior.
- Default:
compare - Compare genomes:
command=compare - Dereplicate genomes:
command=dereplicate
Examples
Compare genomes with default thresholds
analysis=custom
command=compare
2026-SEQ-0001
2026-SEQ-0002
2026-SEQ-0003
Compare a broader primary cluster with FastANI
analysis=custom
mash_ANI_threshold=0.8
comparisonalgorithm=fastANI
2026-SEQ-0001
2026-SEQ-0002
2026-SEQ-0003
Dereplicate a genome set
analysis=custom
command=dereplicate
S_ANI=0.99
coverage_threshold=0.1
2026-SEQ-0001
2026-SEQ-0002
2026-SEQ-0003
The supplied documentation does not provide a working Redmine issue link for dRep.
Interpreting results
When dRep finishes, it uploads a ZIP archive to Dropbox. The archive contains dRep tables, figures, clustering data, and other workflow outputs.
ANI values are calculated only for genomes placed in primary clusters meeting mash_ANI_threshold. If two genomes do not have a secondary ANI value, they may have failed the primary threshold rather than failed processing.
dRep does not directly output a tree file in this workflow. A Mash-based dendrogram can be reproduced by applying hierarchical clustering to the Mash-distance table with the selected clusteralgorithm, but that reconstruction is not equivalent to receiving a maintained tree artifact from the automator.
For command=dereplicate, review the clustering thresholds, coverage requirement, and selected representative genomes. Dereplication results depend directly on these settings and should not be interpreted without recording them.
How long does it take?
Runtime depends on sample count, genome relatedness, thresholds, and the selected secondary algorithm.
A set of 10 or fewer genomes can finish in a few minutes. Larger sets can take several hours. Closely related genomes can also take longer because more pairs advance from Mash primary clustering to the slower secondary ANI comparison.
What can go wrong?
A requested SEQID is unavailable
Symptom: The issue warns that one or more assemblies cannot be found.
Likely cause: An identifier is incorrect or its assembly is unavailable.
What to do: Verify every SEQID and confirm that its assembly exists.
The analysis type is missing or unsupported
Symptom: The automator rejects the request before dRep starts.
Likely cause: analysis=custom is absent or another analysis type was requested.
What to do: Put analysis=custom on the first line.
Expected ANI values are absent
Symptom: Some genome pairs have no secondary ANI result.
Likely cause: They did not cluster together at or above mash_ANI_threshold or did not meet the coverage requirement.
What to do: Review primary clustering and overlap before changing thresholds. Lower thresholds only when methodologically justified.
The comparison or clustering algorithm is unsupported
Symptom: The issue reports an invalid comparisonalgorithm or clusteralgorithm.
Likely cause: The value is misspelled or not in the supported list.
What to do: Copy a documented value exactly.
Dereplication removes more genomes than expected
Symptom: The representative set is smaller or grouped differently than anticipated.
Likely cause: S_ANI, coverage_threshold, primary clustering, or algorithm settings are more permissive or restrictive than intended.
What to do: Review all thresholds and clustering outputs before accepting the representative set.
Related automators
- DiversiTree — selects strains representing diversity within a supplied set rather than dereplicating by ANI.
- CloseRelatives — finds genomes in the CFIA collection closest to one query by Mash distance.
- MashTree — builds a Mash-distance tree for a supplied genome set.
- Unknown Isolate — uses dRep compare as part of a multi-method organism-identification workflow.