Skip to content

Correlate

🚧 Under Construction

This command is under active development and behavior or outputs may change.

Summary

Find co-occurring contigs across samples

Description

Group contigs by cross-sample association patterns.

Supports correlation-based edges, co-occurrence-based edges, or both,
with prevalence and shared-sample thresholds to reduce spurious links.

Writes pairwise edge tables and connected-component style groups using the
selected output prefix.

Usage

rolypoly correlate [OPTIONS]

Options

  • -i, --input: Input table (contig IDs x sample IDs) with presence/absence or abundance values (type: PATH; required; default: Sentinel.UNSET)
  • -o, --output-prefix: Output file prefix (type: TEXT; default: /clusterfs/jgi/scratch/science/metagen/neri/code/rolypoly/correlate)
  • -m, --mode: Analysis mode (type: CHOICE; default: both)
  • --method: Correlation method used in correlation mode (type: CHOICE; default: spearman)
  • --table-type: Input value type (type: CHOICE; default: auto)
  • --min-prevalence: Minimum fraction of samples where contig must be present (type: FLOAT; default: 0.1)
  • --min-correlation: Minimum correlation threshold for keeping contig pairs (type: FLOAT; default: 0.5)
  • --min-shared-samples: Minimum number of shared-present samples for permissive co-occurrence (type: INTEGER; default: 1)
  • --presence-threshold: Values greater than this threshold count as present (type: FLOAT; default: 0.0)
  • --separator: Input delimiter (type: CHOICE; default: auto)
  • -t, --threads: Number of threads (type: INTEGER; default: 1)
  • -g, --log-file: Path to log file (type: TEXT; default: /clusterfs/jgi/scratch/science/metagen/neri/code/rolypoly/correlate_logfile.txt)