Skip to content

Generate Callable Loci

This guide covers how to use clam loci to identify genomic regions with sufficient sequencing depth to be considered callable.

Prerequisites

  • Depth files in one of the supported formats:
    • Per-sample D4 files (uncompressed or bgzipped with index)
    • Merged D4 files (uncompressed only)
    • GVCF files (bgzipped and tabix indexed)
    • A Zarr store from clam collect
  • Optional: population file if analyzing multiple populations

Input Formats

D4 Files

D4 files contain per-base depth estimates. You can generate them from BAM files using mosdepth:

# Generate D4 from BAM
mosdepth --d4 sample sample.bam

# Compress and index (required by clam)
bgzip sample.per-base.d4
bgzip --index sample.per-base.d4.gz

GVCF Files

GVCF files must be bgzipped and tabix indexed:

bgzip sample.g.vcf
tabix -p vcf sample.g.vcf.gz

Basic Usage

clam loci -o callable.zarr -m 10 sample1.d4.gz sample2.d4.gz sample3.d4.gz

This identifies sites where each sample has at least 10x depth.

Setting Depth Thresholds

Per-Sample Thresholds

Control which sites are callable at the individual sample level:

-m, --min-depth
Minimum depth required. Sites below this are not callable.
-M, --max-depth
Maximum depth allowed. Sites above this are not callable (often repetitive regions).
# Require 10-100x depth per sample
clam loci -o callable.zarr -m 10 -M 100 *.d4.gz

Site-Level Thresholds

Apply filters across all samples at each site:

-d, --min-proportion
Fraction of samples that must pass thresholds (0.0-1.0).
--min-mean-depth
Minimum mean depth across samples.
--max-mean-depth
Maximum mean depth across samples.
# Require 80% of samples to be callable at each site
clam loci -o callable.zarr -m 10 -d 0.8 *.d4.gz

Per-Chromosome Thresholds

For sex chromosomes or organellar genomes, use a thresholds file:

# thresholds.tsv
chr1    10  100
chr2    10  100
chrX    5   50
chrY    5   50
chrM    100 10000

clam loci -o callable.zarr --thresholds-file thresholds.tsv *.d4.gz

See Input Formats for file format details.

Specifying Populations

To calculate per-population callable counts (required for dxy and FST downstream):

clam loci -o callable.zarr -m 10 -p populations.tsv *.d4.gz

The population file maps samples to populations:

sample1 PopA
sample2 PopA
sample3 PopB
sample4 PopB

Sample Names

Sample names must exactly match the identifiers in your input files. For D4 files, this is typically the filename prefix (e.g., sample1 for sample1.d4.gz).

See Input Formats for details.

Filtering Chromosomes

Exclude Specific Chromosomes

# Inline
clam loci -o callable.zarr -m 10 -x chrM,chrY *.d4.gz

# From file
clam loci -o callable.zarr -m 10 --exclude-file exclude.txt *.d4.gz

Include Only Specific Chromosomes

# Inline
clam loci -o callable.zarr -m 10 -i chr1,chr2,chr3 *.d4.gz

# From file
clam loci -o callable.zarr -m 10 --include-file autosomes.txt *.d4.gz

Output Options

Per-Sample Masks

By default, clam loci outputs callable counts per population. To output per-sample boolean masks instead:

clam loci -o callable.zarr -m 10 --per-sample *.d4.gz

This is useful when you need sample-level callability information for other analyses.

Using GVCF Input

For GVCF files, you can filter by genotype quality (GQ):

clam loci -o callable.zarr -m 10 --min-gq 20 *.g.vcf.gz

Performance

Use multiple threads for faster processing:

clam loci -o callable.zarr -m 10 -t 16 *.d4.gz

Adjust chunk size if needed (default 1Mb):

clam loci -o callable.zarr -m 10 --chunk-size 500000 *.d4.gz

Complete Example

# Process 50 samples with populations, excluding sex chromosomes
clam loci \
  -o callable.zarr \
  -m 10 \
  -M 100 \
  -d 0.8 \
  -p populations.tsv \
  -x chrX,chrY,chrM \
  -t 16 \
  *.d4.gz

Next Steps

Use the output with clam stat to calculate population genetic statistics:

clam stat -o results/ -w 10000 -c callable.zarr variants.vcf.gz

See Calculate Statistics for details.