mmseqs2_clustering: Recluster sequences of a phyloseq object or cluster a list of...

In MiscMetabar: Miscellaneous Functions for Metabarcoding Analysis

Recluster sequences of a phyloseq object or cluster a list of DNA sequences using MMseqs2 software

Description

A wrapper of the MMseqs2 easy-cluster or easy-linclust workflow.

Usage

mmseqs2_clustering(
  physeq = NULL,
  dna_seq = NULL,
  nproc = 1,
  id = 0.97,
  mmseqs2path = find_mmseqs2(),
  tax_adjust = 0,
  rank_propagation = FALSE,
  mmseqs2_cluster_method = c("easy-cluster", "easy-linclust"),
  coverage = 0.8,
  cov_mode = 0,
  cluster_mode = 0,
  mmseqs2_args = "",
  keep_temporary_files = FALSE
)

Arguments

physeq	(required) a phyloseq-class object obtained using the phyloseq package.
dna_seq	You may directly use a character vector of DNA sequences in place of physeq args. When physeq is set, dna sequences take the value of physeq@refseq.
nproc	(default: 1) Number of threads.
id	(default: 0.97) Minimum sequence identity threshold (0–1).
mmseqs2path	Path to the mmseqs binary (default: find_mmseqs2()).
tax_adjust	(Default 0) See the man page of merge_taxa_vec() for more details. To conserve the taxonomic rank of the most abundant taxa (ASV, OTU, ...), set tax_adjust to 0 (default).
rank_propagation	(logical, default FALSE). Do we propagate the NA value from lower taxonomic rank to upper rank? See the man page of merge_taxa_vec() for more details.
mmseqs2_cluster_method	(default: "easy-cluster") Either "easy-cluster" (cascaded clustering, more sensitive) or "easy-linclust" (linear-time clustering, faster for huge datasets).
coverage	(numeric, default: 0.8) Alignment coverage threshold (0–1), passed to -c.
cov_mode	(integer, default: 0) Coverage mode: 0: alnRes / max(qLen, tLen) 1: alnRes / tLen 2: alnRes / qLen
cluster_mode	(integer, default: 0) Clustering algorithm: 0: greedy set cover (default) 1: connected components 2: greedy incremental
mmseqs2_args	(character, default: "") Additional arguments passed to the MMseqs2 clustering command.
keep_temporary_files	(logical, default: FALSE) Keep intermediate files for debugging?

Details

This function is mainly a wrapper of the work of others. Please cite MMseqs2.

Value

A new object of class physeq or a data.frame of cluster membership if dna_seq was used.

Author(s)

Adrien Taudière

References

MMseqs2 can be downloaded from https://github.com/soedinglab/MMseqs2. More information in the associated publication \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1038/nbt.3988")}.

See Also

postcluster_pq(), vsearch_clustering(), swarm_clustering()

Examples

d_mm <- mmseqs2_clustering(data_fungi_mini)

MiscMetabar documentation built on June 8, 2026, 5:07 p.m.