Skip to content

23-clinical-filter.md

Latest commit

 

History

History
105 lines (70 loc) · 4.13 KB

23-clinical-filter.md

File metadata and controls

105 lines (70 loc) · 4.13 KB

Step 23: Clinical Variant Filter

What This Does

Extracts the small subset of clinically interesting variants from your VEP-annotated VCF. Instead of manually searching through 4-5 million variants, this step produces a focused list of ~200-500 variants that are rare AND functionally impactful.

Why

The biggest challenge after running VEP annotation is: "I have millions of variants, what do I look at?" This step solves that by applying conservative filters to surface the variants most likely to be medically relevant.

Tool

bcftools + bcftools +split-vep plugin (parses VEP CSQ fields structurally — no grep)

Docker Image

staphb/bcftools:1.21

Input

  • VEP-annotated VCF from step 13: ${GENOME_DIR}/${SAMPLE}/vep/${SAMPLE}_vep.vcf (or .vcf.gz)

Command

./scripts/23-clinical-filter.sh your_name

What Gets Filtered

The script produces up to three variant sets (depending on VEP annotations available) that are merged:

HIGH Impact Variants

  • Stop-gained (premature stop codon — breaks the protein)
  • Frameshift insertions/deletions (shifts reading frame — breaks the protein)
  • Splice donor/acceptor (disrupts splicing — breaks the protein)
  • Start-lost (no translation initiation)

Expected count: 100-200 per genome. Uses bcftools +split-vep -s worst to select the most severe consequence per variant.

Rare MODERATE Impact Variants

  • Missense variants (amino acid change) with gnomAD allele frequency < 1%
  • In-frame insertions/deletions with gnomAD AF < 1%

Expected count: 200-400 per genome after frequency filtering. If VEP output lacks gnomAD frequencies (--af_gnomade), all MODERATE variants are included.

ClinVar Pathogenic/Likely Pathogenic (Conditional)

  • Variants with CLIN_SIG containing "pathogenic" (covers both pathogenic and likely_pathogenic)
  • Only runs if VEP was run with --everything or --check_existing (which populates the CLIN_SIG field)
  • Does not use -s worst — a variant is included if ANY transcript annotation has a pathogenic ClinVar entry

Expected count: 10-50 per genome (depends on ClinVar version).

Output

File Contents Size
${SAMPLE}_clinical.vcf.gz Combined clinically interesting VCF < 5 MB
${SAMPLE}_clinical_summary.tsv Human-readable tab-delimited table < 1 MB
${SAMPLE}_high_impact.vcf.gz HIGH impact variants only < 2 MB
${SAMPLE}_rare_moderate.vcf.gz Rare MODERATE variants only < 3 MB
${SAMPLE}_clinvar_pathogenic.vcf.gz ClinVar P/LP only (if CLIN_SIG available) < 1 MB

Runtime

~5-10 minutes (I/O-bound, reading the large VEP VCF)

How to Use the Output

Quick look at the summary

# View the most important variants
column -t ${GENOME_DIR}/${SAMPLE}/clinical/${SAMPLE}_clinical_summary.tsv | head -20

Cross-reference with ClinVar

# Find which clinical variants are also in ClinVar
docker run --rm -v "${GENOME_DIR}:/genome" staphb/bcftools:1.21 \
  bcftools isec -n=2 -w1 \
    /genome/${SAMPLE}/clinical/${SAMPLE}_clinical.vcf.gz \
    /genome/clinvar/clinvar.vcf.gz \
    -Oz -o /genome/${SAMPLE}/clinical/${SAMPLE}_clinical_clinvar.vcf.gz

Load in a genome browser

The _clinical.vcf.gz file is small enough to load in IGV Web or gene.iobio for visual inspection.

Limitations

  • This is a computational filter, not a clinical interpretation
  • Some pathogenic variants are LOW impact (e.g., synonymous variants affecting splicing, regulatory variants) and will be missed by this filter
  • gnomAD frequency filtering depends on VEP having annotated the gnomAD fields correctly
  • Always cross-reference findings with ClinVar and consult a professional for clinical decisions

Notes

  • No additional Docker images required — uses the same bcftools image as other steps
  • Uses bcftools +split-vep to parse VEP's pipe-delimited CSQ annotation structurally (not grep)
  • The VEP VCF is compressed and indexed automatically if needed
  • PASS filter is applied to exclude low-quality variant calls
  • Available CSQ subfields are auto-detected — ClinVar and gnomAD filters are skipped gracefully if not present