cluster¶

Sequence clustering wrapper around the external vclust tool with three preconfigured modes for common viral genomics workflows.

Overview¶

The phu cluster command provides a simplified interface to vclust, implementing common use cases (see the vclust wiki) while allowing advanced customization via --vclust-params.

Synopsis¶

phu cluster --mode <MODE> --input-contigs <FASTA_FILE> [OPTIONS]

Outputs are placed in an output folder (default clustered-contigs/) and include cluster assignment TSVs and representative FASTA files, for example:

clustered-contigs/
├── ani.ids.tsv
├── ani.tsv
├── cluster_representatives_ids.txt
├── fltr.txt
├── representatives.fna
└── species.tsv

Modes¶

dereplication — remove redundant sequences while keeping representatives (cd-hit/ANI-based).
votu — cluster into viral Operational Taxonomic Units (leiden/ANI-based) following MIUViG-style defaults.
species — classify sequences into species (complete/TANI-based) following ICTV-style defaults.

Default parameters by mode¶

Parameter	dereplication	votu	species
Algorithm	cd-hit	leiden	complete
Metric	ani	ani	tani
ANI cutoff	95%	95%	95%
Query coverage	85%	85%	None
Pre-filter min-ident	95%	95%	70%

Command Options¶

 Sequence clustering wrapper around external 'vclust' with three modes.      

 For advanced usage, provide custom vclust parameters as a quoted string.    
 See the vclust wiki for parameter details:                                  
 https://github.com/refresh-bio/vclust/wiki                                  

 Example:                                                                    
     phu cluster --mode votu --input-contigs genomes.fna                                               

╭─ Options ─────────────────────────────────────────────────────────────────╮
│ *  --mode                   [dereplication|votu|s  dereplication | votu | │
│                             pecies]                species                │
│                                                    [required]             │
│ *  --input-contigs          PATH                   Input FASTA [required] │
│    --output-folder          PATH                   Output directory       │
│                                                    [default:              │
│                                                    clustered-contigs]     │
│    --threads                INTEGER RANGE [x>=0]   0=all cores; otherwise │
│                                                    N threads              │
│                                                    [default: 0]           │
│    --vclust-params          TEXT                   Custom vclust          │
│                                                    parameters:            │
│                                                    "--min-kmers 20        │
│                                                    --outfmt lite --ani    │
│                                                    0.97"                  │
│    --help           -h                             Show this message and  │
│                                                    exit.                  │
╰───────────────────────────────────────────────────────────────────────────╯

Examples¶

# Dereplicate viral contigs (default outputs in clustered-contigs/)
phu cluster --mode dereplication --input-contigs viral_contigs.fna

# Cluster into vOTUs following MIUViG standards
phu cluster --mode votu --input-contigs viral_contigs.fna

# Species classification following ICTV standards
phu cluster --mode species --input-contigs complete_genomes.fna

Advanced: pass custom vclust parameters¶

phu cluster --mode species --input-contigs genomes.fna \
  --vclust-params="--metric tani --tani 0.70"

Treat species clustering as genus-level by lowering similarity to 70%

Notes¶

--vclust-params provides full control over vclust parameters; use it to tune behavior for large or highly-redundant datasets.
For reproducibility, save the full vclust command and parameters used (the tool often logs them in the output folder).