TinJasmine

TinJasmine is a CWL implementation of a germline variant calling pipeline. It is modeled on the Huang et al. Pathogenic Germline Variants in 10,389 Adult Cancers

See CromwellRunner for information about managing pipeline runs using Cromwell on LSF environments

Versions

• v1.4 - Filtering updates
  • Removing spanning deletions (ALT=*) from GATK calls
  • Add post-merge normalization step to split multiallelic variants which were re-combined in merge step
  • Filtering for MLEAF/MLEAC issue
  • Bcftools moved to separate submodule and updated to v1.10.2
• v1.3 - Incorporates bcftools reheader to make all VCF headers consistent
• v1.2
  • VLD filter which has parameters passed via CWL
  • Updated VEP annotation with options for v99, v100, and v102
• v1.1
  • Initial version of TinJasmine compatible with CromwellRunner

Subprojects

The TinJasmine CWL workflow consists of number of smaller CWL tools which are developed as independent projects:

• GATK_GermlineCaller
• Varscan_GermlineCaller
• Pindel_GermlineCaller
• varscan_vcf_remap
• MergeFilterVCF
• TinDaisy-VEP
• HotspotFilter
• VCF2MAF
• VLD_Filter
• bcftools
  • Wrapper around bcftools 1.10.2 with related filter scripts

Installation

All of the subprojects can be installed together with TinJasmine as,

git clone --recurse-submodules https://github.com/ding-lab/TinJasmine.git

Output

Three output files are produced:

• all_call_vcf: Contains all pre-merge PASS variants. Any filters applied after merge step will add a value to FILTER field but variant will be retained
  • Filename: output_vep.vcf
• clean_VC: Contains only those variants which have FILTER value PASS after all post-merge filters applied
  • Filename: HotspotFiltered.vcf
• clean_MAF: MAF version of all variants found in clean VCF
  • Filename: result.maf

Contact info

m.wyczalkowski@wustl.edu