Mmseq
Sequence-identity clustering with MMseqs2 easy-cluster.
MMseqsClustering(params)
¶
Bases: Clusterer[MMseqsParams]
MMseqs2 easy-cluster sequence-identity clustering.
Initialize the MMseqs2 clusterer.
| PARAMETER | DESCRIPTION |
|---|---|
params
|
the
TYPE:
|
cluster(sequences)
¶
Cluster sequences by sequence identity with MMseqs2.
| PARAMETER | DESCRIPTION |
|---|---|
sequences
|
a dictionary with the keys being stable id of the sequence (used for fasta headers) and the values being the sequence string.
TYPE:
|
| RETURNS | DESCRIPTION |
|---|---|
ClusterResult
|
|
ClusterResult
|
keyed by the keys in |
cluster_fasta(*fasta_paths)
¶
Cluster sequences already on disk as FASTA file(s) instead of in-memory dictionary.
| PARAMETER | DESCRIPTION |
|---|---|
*fasta_paths
|
One or more paths to FASTA files.
TYPE:
|
| RETURNS | DESCRIPTION |
|---|---|
ClusterResult
|
A |
ClusterResult
|
them, i.e. truncated at the first whitespace). |
from_cluster_tsv(tsv_path, params)
¶
Build a ClusterResult object from an existing MMseqs2 *_cluster.tsv result.
| PARAMETER | DESCRIPTION |
|---|---|
tsv_path
|
Path to an MMseqs2 cluster TSV.
TYPE:
|
params
|
The params used to produce it (recorded on the result).
TYPE:
|
| RETURNS | DESCRIPTION |
|---|---|
ClusterResult
|
A |
MMseqsParams(min_seq_id=0.9, sensitivity=7.0, coverage_mode=0, cluster_mode=0, self_match=True, cluster_reassign=True, sort_results=True, output_dir=None)
¶
Bases: ClusterParams
Parameters for MMseqs2 sequence-identity clustering.
cluster_mode = 0
¶
Cluster mode (--cluster-mode), one of {0, 1, 2, 3}.
cluster_reassign = True
¶
Whether to reassign members to their best cluster after cascaded clustering
(--cluster-reassign). Kept as a param because it changes cluster assignment;
if you never toggle it, it could instead be an internal constant.
coverage_mode = 0
¶
Coverage mode (--cov-mode), one of {0, 1, 2, 3}.
method = ClusterMethod.MMSEQS2
¶
The clustering method used, which is MMseqs2.
min_seq_id = 0.9
¶
Minimum sequence identity for two sequences to cluster together (--min-seq-id).
Must be between 0 and 1; higher values are stricter.
output_dir = None
¶
Directory to write MMseqs2 outputs to. Default is None, which uses a temporary directory that is deleted after clustering. If a path is given, outputs are written there and kept after the run.
self_match = True
¶
Whether to add self-matches (--add-self-matches).
sensitivity = 7.0
¶
Search sensitivity (-s). Higher values find more remote homologs at
higher cost. MMseqs2's scale runs 1.0 (faster), 4.0 (fast), to 7.5
(sensitive); accepted range is 1.0–7.5.
sort_results = True
¶
Whether to sort alignment results by sequence identity (--sort-results).
It sets the TSV row order that determines integer cluster-id assignment. Defaults is True.