ngs checkmate workflow

Introduction

Based on the tool from https://github.com/parklab/NGSCheckMate, "NGSCheckMate uses depth-dependent correlation models of allele fractions of known single-nucleotide polymorphisms (SNPs) to identify samples from the same individual." Contains preprocessing tools and workflows, as well as a workflow for batch processing.

Workflows

References obtainable from https://cavatica.sbgenomics.com/u/kfdrc-harmonization/kf-references

bcf_call.cwl

Creates input vcfs for ngs checkmate. Especially useful to run when inputs are large WGS bam files.

inputs

inputs:
  input_align: File[]
  chr_list: File
  reference_fasta: File
  snp_bed: File

Suggested inputs:

chr_list: chr_list.txt
snp_bed: SNP_hg38_liftover_wChr.bed
reference_fasta: Homo_sapiens_assembly38.fasta

outputs

bcf_called_vcf: {type: File[], outputSource: bcf_filter/bcf_call}

ngs_checkmate_wf.cwl

Runs ngscheckmate in vcf mode - requires output from bcf_call step.

inputs

inputs:
  input_vcf:
    type:
        type: array
        items:
            type: array
            items: File
  
  snp_bed: File
  output_basename: string[]
  ram: 
    type: ['null', int]
    default: 4000

1. Ram in megabytes, input param optional - use if you plan on batching ~20+ vcfs.
2. input_vcf is an array of arrays - basically an array of groups of vcfs that you'd like ot see evaluated together.
3. output_basename is an array of file output prefixes - should line up with the first level of array elements from input_vcf

outputs

outputs:
  match_results: {type: 'File[]', outputSource: ngs_checkmate/match_results}
  correlation_matrix: {type: 'File[]', outputSource: ngs_checkmate/correlation_matrix}

Tools:

bcf_filter.cwl

Used by bcf_call.cwl, can take cram or bam as input.

inputs

inputs:
  input_align:
    type: File
    secondaryFiles: |
        ${
          if (inputs.input_align.nameext == '.cram'){
            return inputs.input_align.basename + '.crai';
          }
        else {
          return inputs.input_align.nameroot + '.bai';
        }
        }
  chr_list: File
  reference_fasta:
    type: File
    secondaryFiles: ['.fai']
  snp_bed: File

outputs

outputs:
  bcf_call:
    type: File
    outputBinding:
      glob: "$(inputs.input_align.nameroot).bcf.called.vcf"

ngs_checkmate_vcf.cwl

Takes an array of vcfs in which bams have been filtered using bcf tools and outputs match results and a correlation matrix.

inputs

inputs:
  input_vcf:
    type: File[]
  snp_bed: File
  output_basename: string
  ram:
    type: ['null', int]
    default: 4000

outputs

outputs:
  match_results:
    type: File
    outputBinding:
      glob: "$(inputs.output_basename)_all.txt"

  correlation_matrix:
    type: File
    outputBinding:
      glob: "$(inputs.output_basename)_corr_matrix.txt "

rna_tx2genome.cwl

In the rare event that your input is a transcriptome bam, this will convert to the necessary sorted genome bam.

inputs

inputs:
  input_bam: {type: File, doc: "Input transcriptome bam"}
  genomeDir: {type: File, doc: "rsem tar gzipped reference"}

outputs

outputs:
  genome_bam:
    type: File
    outputBinding:
      glob: "$(inputs.input_bam.nameroot).converted.bam"
    secondaryFiles: [^.bai]

samtools_sort_index.cwl

In the event that your bam is not coord sorted, use this tool first

inputs

inputs:
  input_align: File

outputs

outputs:
  sorted_bam:
    type: File
    outputBinding:
      glob: "$(inputs.input_align.nameroot).sorted.bam"
  sorted_bai:
    type: File
    outputBinding:
      glob: "$(inputs.input_align.nameroot).sorted.bai"