eggd vcf handler for uranus

What does this app do?

This app uses bedtools, bcftools and VEP to take VCFs from sentieon mutect2 and pindel:

• Annotates and filters sentieon mutect2 and pindel VCFs
• Produces an excel workbook with sub-panels presented in separate sheets; also provides preformatted text to aid Epic data entry
• Produces a VCF that can be used as input for BSVI, as BSVI can't handle mutect2's (VCFv4.2 compliant) representation of multiallelic variants in VCF

What are typical use cases for this app?

This app uses the follow tools which are app assets:

• bcftools (v1.12)
• bedtools (v2.30.0)
• python_packages (numpy-1.20.1, pandas-1.2.3, pytz-2021.1, XlsxWriter-1.4.0)

This app uses the following provided as inputs:

• VEP (v103.1) (docker image)
• VEP refseq (v103) annotation sources
• CADD (v1.6) which now includes splicing
• ClinVar VCF (20210501 release) modified to add chr prefix
• Merged VCF containing counts of each variant detected known set of samples (provided as separate input from tarball)

This app has access to the Internet

What data are required for this app to run?

• VCF output from sentieon mutect2 as part of Uranus workflow
• VCF output (+ index) from cgppindel as part of Uranus workflow
• BED file that details ROIs for myeloid NGS panel
• BED file that details ROIs for filtering cgppindel output
• Genome FASTA and index that was used by to generate the VCF
• VEP docker image (vep_docker)
• VEP plugins (vep_plugins)
• VEP reference files (vep_refs)
• VEP annotation sources (vep_annotation)
  • Cosmic Coding Variants VCF
  • Cosmic NonCoding Variants VCF
  • ClinVar VCF
  • CADD
• MAF file created from known set of samples

What does this app output?

• Excel workbook of annotated variants
• BSVI-friendly VCF
• Intermediate VCFs
• header modified VCFs for uploading to openCGA

How does this app work?

• Filters mutect2 VCF with bedtools:
  • retain variants within ROI
• Filters mutect2 VCF with bcftools:
  • retain positions where at least one variant has AF > 0.03
  • retain positions where DP >99
  • split multiallelics using --keep-sum AD which changes the ref AD to be the sum of AD's
  • split multiallelics requires fixing AD and RPA number field in header from . to R
• Filters cgppindel VCF with bcftools:
  • only indels that intersect with the exons of interest bed file
  • only insertions with length greater than 2. This will remove the 1 bp false positive insertions.
• Annotates mutect2 and cgppindel VCF with VEP:
  • Annotate against specified RefSeq transcripts with:
    • gene symbol
    • variant class
    • variant consequence
    • exon number
    • HGVS c. & p.
    • gnomAD AF
    • SIFT
    • PolyPhen
    • dbSNP
    • COSMIC
    • ClinVar
    • CADD
    • previous counts
• Filters VCFs with VEP:
  • Retain variants with gnomAD AF < 0.1
  • Remove synonymous variants
  • Filters by transcripts defined in code.sh for each panel
• Generates variant lists (one panel + pindel VCF per sheet) in an excel workbook
• Generates BSVI-friendly VCF

Limitations

• Designed to be used as part of Uranus workflow for processing myeloid NGS panel
• When building the app, mark-section and mark-success must be executable for the app to build correctly - these two scripts are executable in this repo but it can be easy to miss if the executable bit is lost as it's a metadata change