Step 16: Coverage QC and Sex Chromosome Verification¶

What This Does¶

Ultra-fast whole-genome coverage profiling directly from the BAM index file. Infers sex chromosome copy number (CNchrX, CNchrY) and detects sex chromosome aneuploidies (XXY, XYY, X0). Produces per-chromosome depth uniformity plots.

Why¶

Coverage QC catches alignment problems, sample swaps, and sequencing artifacts early — before spending hours on variant calling. Comparing the sex inferred from X/Y coverage with the sex you declare is a cheap sample-identity check: a swapped sample usually shows up as a mismatch. It can also reveal sex-chromosome aneuploidies like Klinefelter syndrome (XXY). It cannot tell two samples of the same sex apart; step 33 checks the sex again from the reads, compares the samples of a run with each other, and estimates contamination.

Tool¶

goleft indexcov (Brent Pedersen)

Docker Image¶

GOLEFT_IMAGE

Pinned in versions.env; Image versions lists the current tag.

Command¶

./scripts/16-indexcov.sh your_name male     # or female; run-all.sh passes the sex you give it
./scripts/16-indexcov.sh your_name          # no declared sex: report only, no check

The script runs:

source versions.env   # from the repository root
docker run --rm \
  --cpus 1 --memory 1g \
  -v ${GENOME_DIR}:/genome \
  "${GOLEFT_IMAGE}" \
  goleft indexcov \
  --directory /genome/${SAMPLE}/indexcov \
  /genome/${SAMPLE}/aligned/${SAMPLE}_sorted.bam

Sex check¶

goleft writes the inferred sex to indexcov-indexcov.ped, whose columns are #family_id sample_id paternal_id maternal_id sex phenotype CNchrX CNchrY .... The sex column uses PED coding: 1 male, 2 female, anything else unknown. The phenotype column is always -9. The script finds the columns by header name, prints the raw row, CNchrX, CNchrY and the inferred sex, and then compares it with the sex you declared:

Match: prints Sex check: OK.
Mismatch: prints both values and exits non-zero. A mismatch means a sample swap, a wrong declared sex, or a sex-chromosome aneuploidy. Steps that take the declared sex (ExpansionHunter, step 9) would otherwise use a wrong value without warning.
SEX_CHECK=warn ./scripts/16-indexcov.sh your_name female prints the mismatch and exits 0, for when you know why they differ (for example 47,XXY).

Output Files¶

File	Description
`indexcov-indexcov.ped`	PED file with CN values for chrX, chrY, and autosomes
`indexcov-indexcov.roc`	ROC-like data for each chromosome
`indexcov-indexcov.bed.gz`	Per-16KB-bin normalized depth across all chromosomes
`index.html`	Interactive HTML report with all plots

Interpretation¶

Sex Chromosome Copy Number¶

Karyotype	CNchrX	CNchrY	Meaning
46,XY (male)	~1.0	~1.0	Normal male
46,XX (female)	~2.0	~0.0	Normal female
47,XXY (Klinefelter)	~2.0	~1.0	Male with extra X
47,XYY	~1.0	~2.0	Male with extra Y
45,X (Turner)	~1.0	~0.0	Female with single X

Coverage Uniformity¶

Flat depth across a chromosome = good sequencing
Dips or spikes = possible CNVs, GC bias, or capture artifacts
Systematic deviations across all chromosomes = library prep or sequencing problems

Runtime¶

~5 seconds per sample.

Notes¶

Should be run as an early QC step after alignment (step 2). It only reads the .bai index, not the full BAM.
Requires the BAM index (.bai) to exist alongside the BAM file.
Works on any number of samples simultaneously — useful for batch QC.
The HTML report opens in any browser, but it loads Chart.js and jQuery from public CDNs; offline, the page opens without its plots.
For single-sample runs, the sex chromosome plot is still useful but the population-level clustering view is less informative.