Arguments

All arguments are listed below. wisp --help shows the same information.

Commands

wisp has two subcommands:

from-alignments: Build a population count mask from BAM/CRAM files via mosdepth.
from-vcf: Build a population count mask from a prefiltered all-sites VCF with per-sample FORMAT/DP values.

Core arguments

--min-dp INTEGER: Minimum depth for a sample to pass a site. Must be non-negative. Required unless from-alignments is run with --variants-vcf and the VCF contains per-sample FORMAT/DP values for threshold estimation. --min-dp 0 requires --mask because no genome-wide coordinate file is supplied.
--out PATH: Output directory for the final wisp.bed.gz and tabix index.

Input-specific arguments

--samples PATH: Sample metadata TSV for BAM/CRAM mode. Must provide sample ID, population, and alignment path columns. See Input files.
--variants-vcf PATH: Optional variants-only VCF for BAM/CRAM mode. When present, omitted --min-dp and --max-dp values are estimated from positive per-sample FORMAT/DP values among selected samples and omitted --min-mapq is estimated from INFO/MQ. Manual threshold flags override these estimates. Indels, symbolic structural variants, breakends, and multi-nucleotide polymorphisms are converted to exclusion intervals and removed from all sample pass BEDs, so those spans are omitted from the sparse final BED and interpreted as zero passing samples in every population.
--all-sites-vcf PATH: All-sites VCF for VCF mode. Must include per-sample FORMAT/DP values for every sample in --popfile.
--popfile PATH: Sample/population TSV for VCF mode. Required with --all-sites-vcf.

Shared optional arguments

--work PATH: Working directory for intermediate files. Defaults to <out>/work.
--mask PATH: Restrict output to sites within this BED. In BAM/CRAM mode, the mask is normalized to three columns, sorted, and merged before sample pass intervals are clipped. In VCF mode, only bases within the mask intervals are emitted.
--output-prefix TEXT: Output filename stem within --out. Defaults to wisp, producing wisp.bed.gz and wisp.bed.gz.tbi. .bed.gz is always appended.
--keep-work: Keep intermediate files in the working directory. By default, work files are removed after a successful run.
--force: Overwrite existing final outputs. Without this flag, wisp refuses to replace wisp.bed.gz or wisp.bed.gz.tbi.
--version: Print the installed wisp version and exit.
--help: Print the full help message and exit.

BAM/CRAM mode options

--threads INTEGER: mosdepth threads per sample. Defaults to 1.
--jobs INTEGER: Samples to process concurrently. Defaults to 1. Total parallelism is --jobs × --threads.
--min-mapq INTEGER: Minimum read mapping quality. In BAM/CRAM mode, defaults from --variants-vcf when available.
--max-dp INTEGER: Maximum depth to pass a site. In VCF mode, omitted means no upper depth bound. In BAM/CRAM mode, defaults from --variants-vcf when available.
--exclude-flag INTEGER: SAM FLAG bits to exclude reads.
--include-flag INTEGER: SAM FLAG bits required to include reads. Passed through to mosdepth --include-flag.
--reference PATH: FASTA reference for CRAM inputs.
--fast-mode: Opt into mosdepth --fast-mode. Strict per-base depth counting is the default, so the former --strict-depth flag is no longer needed. Fast mode is quicker, but allows mosdepth's fast-mode approximations.

Examples

BAM/CRAM mode:

wisp from-alignments \
  --samples tests/test_data/1000g_5sample_chr20_smoke/samples.tsv \
  --min-dp 10 \
  --variants-vcf validation/cohort.variants.vcf.gz \
  --mask tests/test_data/1000g_5sample_chr20_smoke/targets.bed \
  --out results \
  --work work \
  --threads 4 \
  --jobs 2 \
  --keep-work

All-sites VCF mode:

wisp from-vcf \
  --all-sites-vcf validation/cohort.all_sites.vcf.gz \
  --popfile validation/sample_populations.tsv \
  --min-dp 10 \
  --max-dp 80 \
  --mask validation/targets.bed \
  --out results \
  --work work