BroadOnPremMalariaPipelineTasks

VariantRecalibrator

description
Run GATK VariantRecalibrator/ApplyRecalibration (VQSR) over the joint-call VCF for SNPs then indels. The VQSR model-fitting knobs are exposed so small callsets (where the negative model can otherwise find no 'bad' variants) can be tuned at runtime.

Inputs

Required

  • input_vcf (File, required)
  • prefix (String, required)
  • reference_dict (File, required)
  • reference_fai (File, required)
  • reference_fasta (File, required)
  • resource_vcf_3d7_hb3 (File, required)
  • resource_vcf_3d7_hb3_index (File, required)
  • resource_vcf_7g8_gb4 (File, required)
  • resource_vcf_7g8_gb4_index (File, required)
  • resource_vcf_hb3_dd2 (File, required)
  • resource_vcf_hb3_dd2_index (File, required)

Optional

  • runtime_attr_override (RuntimeAttr?): Override default runtime attributes.

Defaults

  • bad_lod_cutoff (Float, default=-5.0): --badLodCutoff for both models (GATK default -5.0). Raise toward 0 (e.g. -2.0) so more variants enter the negative training set when none score below -5.
  • indel_max_gaussians (Int, default=4): Indel model --maxGaussians (default 4).
  • min_num_bad_variants (Int, default=1000): --minNumBadVariants for both models (GATK default 1000). Lower to accept a small negative training set.
  • snp_max_gaussians (Int, default=8): SNP model --maxGaussians (default 8). Lower it for small callsets that cannot support 8 Gaussians.

Outputs

  • vcf (File)

VariantRecalibratorBestEffort

description
Best-effort VQSR: identical GATK VariantRecalibrator/ApplyRecalibration commands as VariantRecalibrator, but attempted non-fatally. On any VQSR failure (typically the INDEL negative model on small callsets: 'No data found') the task deletes partial outputs, reports vqsr_ok=false, and exits 0 so the workflow can end cleanly without a recalibrated VCF.

Inputs

Required

  • input_vcf (File, required)
  • prefix (String, required)
  • reference_dict (File, required)
  • reference_fai (File, required)
  • reference_fasta (File, required)
  • resource_vcf_3d7_hb3 (File, required)
  • resource_vcf_3d7_hb3_index (File, required)
  • resource_vcf_7g8_gb4 (File, required)
  • resource_vcf_7g8_gb4_index (File, required)
  • resource_vcf_hb3_dd2 (File, required)
  • resource_vcf_hb3_dd2_index (File, required)

Optional

  • runtime_attr_override (RuntimeAttr?): Override default runtime attributes.

Defaults

  • bad_lod_cutoff (Float, default=-5.0): --badLodCutoff for both models (GATK default -5.0).
  • indel_max_gaussians (Int, default=4): Indel model --maxGaussians (default 4).
  • min_num_bad_variants (Int, default=1000): --minNumBadVariants for both models (GATK default 1000).
  • snp_max_gaussians (Int, default=8): SNP model --maxGaussians (default 8).

Outputs

  • vqsr_ok (Boolean)
  • vcf (File?)

SortCompressIndexVcf

Inputs

Required

  • input_vcf (File, required)

Optional

  • runtime_attr_override (RuntimeAttr?)

Outputs

  • vcf (File)
  • vcf_index (File)