Merge pull request #20 from kids-first/feature/mb-add-skip-norm-flag

🔨 Feature add skip norm flag
kids-first · Nov 21, 2024 · a972882 · a972882
2 parents 934ea4c + f8e545d
commit a972882
Show file tree

Hide file tree

Showing 9 changed files with 75 additions and 86 deletions.
diff --git a/docs/SOMATIC_SNV_ANNOT_README.md b/docs/SOMATIC_SNV_ANNOT_README.md
@@ -58,21 +58,26 @@ For each input, the sub-bullet refers to when to use the suggested input
    - Mutect2: "gnomad_3_1_1_AC,gnomad_3_1_1_AN,gnomad_3_1_1_AF,gnomad_3_1_1_nhomalt,gnomad_3_1_1_AC_popmax,gnomad_3_1_1_AN_popmax,gnomad_3_1_1_AF_popmax,gnomad_3_1_1_nhomalt_popmax,gnomad_3_1_1_AC_controls_and_biobanks,gnomad_3_1_1_AN_controls_and_biobanks,gnomad_3_1_1_AF_controls_and_biobanks,gnomad_3_1_1_AF_non_cancer,gnomad_3_1_1_primate_ai_score,gnomad_3_1_1_splice_ai_consequence,gnomad_3_1_1_AF_non_cancer_afr,gnomad_3_1_1_AF_non_cancer_ami,gnomad_3_1_1_AF_non_cancer_asj,gnomad_3_1_1_AF_non_cancer_eas,gnomad_3_1_1_AF_non_cancer_fin,gnomad_3_1_1_AF_non_cancer_mid,gnomad_3_1_1_AF_non_cancer_nfe,gnomad_3_1_1_AF_non_cancer_oth,gnomad_3_1_1_AF_non_cancer_raw,gnomad_3_1_1_AF_non_cancer_sas,gnomad_3_1_1_AF_non_cancer_amr,gnomad_3_1_1_AF_non_cancer_popmax,gnomad_3_1_1_AF_non_cancer_all_popmax,gnomad_3_1_1_FILTER,MBQ,TLOD,HotSpotAllele"
    - Lancet: "gnomad_3_1_1_AC,gnomad_3_1_1_AN,gnomad_3_1_1_AF,gnomad_3_1_1_nhomalt,gnomad_3_1_1_AC_popmax,gnomad_3_1_1_AN_popmax,gnomad_3_1_1_AF_popmax,gnomad_3_1_1_nhomalt_popmax,gnomad_3_1_1_AC_controls_and_biobanks,gnomad_3_1_1_AN_controls_and_biobanks,gnomad_3_1_1_AF_controls_and_biobanks,gnomad_3_1_1_AF_non_cancer,gnomad_3_1_1_primate_ai_score,gnomad_3_1_1_splice_ai_consequence,gnomad_3_1_1_AF_non_cancer_afr,gnomad_3_1_1_AF_non_cancer_ami,gnomad_3_1_1_AF_non_cancer_asj,gnomad_3_1_1_AF_non_cancer_eas,gnomad_3_1_1_AF_non_cancer_fin,gnomad_3_1_1_AF_non_cancer_mid,gnomad_3_1_1_AF_non_cancer_nfe,gnomad_3_1_1_AF_non_cancer_oth,gnomad_3_1_1_AF_non_cancer_raw,gnomad_3_1_1_AF_non_cancer_sas,gnomad_3_1_1_AF_non_cancer_amr,gnomad_3_1_1_AF_non_cancer_popmax,gnomad_3_1_1_AF_non_cancer_all_popmax,gnomad_3_1_1_FILTER,MS,FETS,HotSpotAllele"
    - Vardict: "gnomad_3_1_1_AC,gnomad_3_1_1_AN,gnomad_3_1_1_AF,gnomad_3_1_1_nhomalt,gnomad_3_1_1_AC_popmax,gnomad_3_1_1_AN_popmax,gnomad_3_1_1_AF_popmax,gnomad_3_1_1_nhomalt_popmax,gnomad_3_1_1_AC_controls_and_biobanks,gnomad_3_1_1_AN_controls_and_biobanks,gnomad_3_1_1_AF_controls_and_biobanks,gnomad_3_1_1_AF_non_cancer,gnomad_3_1_1_primate_ai_score,gnomad_3_1_1_splice_ai_consequence,gnomad_3_1_1_AF_non_cancer_afr,gnomad_3_1_1_AF_non_cancer_ami,gnomad_3_1_1_AF_non_cancer_asj,gnomad_3_1_1_AF_non_cancer_eas,gnomad_3_1_1_AF_non_cancer_fin,gnomad_3_1_1_AF_non_cancer_mid,gnomad_3_1_1_AF_non_cancer_nfe,gnomad_3_1_1_AF_non_cancer_oth,gnomad_3_1_1_AF_non_cancer_raw,gnomad_3_1_1_AF_non_cancer_sas,gnomad_3_1_1_AF_non_cancer_amr,gnomad_3_1_1_AF_non_cancer_popmax,gnomad_3_1_1_AF_non_cancer_all_popmax,gnomad_3_1_1_FILTER,MSI,MSILEN,SOR,SSF,HotSpotAllele"
+   - Consensus: "gnomad_3_1_1_AC,gnomad_3_1_1_AN,gnomad_3_1_1_AF,gnomad_3_1_1_nhomalt,gnomad_3_1_1_AC_popmax,gnomad_3_1_1_AN_popmax,gnomad_3_1_1_AF_popmax,gnomad_3_1_1_nhomalt_popmax,gnomad_3_1_1_AC_controls_and_biobanks,gnomad_3_1_1_AN_controls_and_biobanks,gnomad_3_1_1_AF_controls_and_biobanks,gnomad_3_1_1_AF_non_cancer,gnomad_3_1_1_primate_ai_score,gnomad_3_1_1_splice_ai_consequence,gnomad_3_1_1_AF_non_cancer_afr,gnomad_3_1_1_AF_non_cancer_ami,gnomad_3_1_1_AF_non_cancer_asj,gnomad_3_1_1_AF_non_cancer_eas,gnomad_3_1_1_AF_non_cancer_fin,gnomad_3_1_1_AF_non_cancer_mid,gnomad_3_1_1_AF_non_cancer_nfe,gnomad_3_1_1_AF_non_cancer_oth,gnomad_3_1_1_AF_non_cancer_raw,gnomad_3_1_1_AF_non_cancer_sas,gnomad_3_1_1_AF_non_cancer_amr,gnomad_3_1_1_AF_non_cancer_popmax,gnomad_3_1_1_AF_non_cancer_all_popmax,gnomad_3_1_1_FILTER,MQ,MQ0,CAL,HotSpotAllele"
  - `retain_ann` # Similar to above, if run for KF harmonization, recommend the following:
    - Strelka2: "HGVSg"
    - Mutect2: "HGVSg"
    - Lancet: "HGVSg"
    - Vardict: "HGVSg"
+   - Consensus: "HGVSg"
  - `bcftools_strip_columns` # if reannotating an old file:
    - "FILTER/GNOMAD_AF_HIGH,FILTER/NORM_DP_LOW,INFO/CSQ,INFO/HotSpotAllele" # recommended if re-annotating from an older VEP cache
    - "FILTER/GNOMAD_AF_HIGH,FILTER/NORM_DP_LOW,INFO/HotSpotAllele" # recommended if repeating hot spot and want to keep VEP
  - `bcftools_prefilter_csv` # if annotating a file with calls you want screen for, use this. i.e `FILTER="PASS"`
+ - `disable_norm` # set to `True` if existing input already normalized or of you have justification for skipping this step
  - `disable_vep_annotation` # set to `True` if existing VEP annotation of file is ok
+ - `disable_hotspot_annotation` # set to `True` if existing HotSpot annotation is ok
  - `tool_name`:
    - `Strelka2`: `strelka2_somatic`
    - `Mutect2`: `mutect2_somatic`
    - `Lancet`: `lancet_somatic`
    - `VarDict Java`: `vardict_somatic`
+   - `consensus`: `consensus_somatic`
  - `vep_cores`: `16`
  - `vep_ram`: `32`
  - `vep_buffer`: `5000`

diff --git a/tools/add_strelka2_fields.cwl b/tools/add_strelka2_fields.cwl
@@ -46,10 +46,10 @@ inputs:
       prefix: '--output_basename'
   cores:
     type: int?
-    default: 4
+    default: 8
   ram:
     type: int?
-    default: 3 
+    default: 16
     doc: 'RAM requirement in GB'
 
 outputs:

diff --git a/tools/bcftools_strip_ann.cwl b/tools/bcftools_strip_ann.cwl
@@ -6,8 +6,8 @@ requirements:
   - class: ShellCommandRequirement
   - class: InlineJavascriptRequirement
   - class: ResourceRequirement
-    ramMin: 8000
-    coresMin: 4
+    ramMin: 16000
+    coresMin: 8
   - class: DockerRequirement
     dockerPull: 'pgc-images.sbgenomics.com/d3b-bixu/vcfutils:latest'
 
@@ -18,7 +18,7 @@ arguments:
     valueFrom: >-
       set -eo pipefail
 
-      (bcftools annotate -x $(inputs.strip_info) $(inputs.input_vcf.path) -O z
+      (bcftools annotate -x $(inputs.strip_info) $(inputs.input_vcf.path) --threads 8 -O z
       -o $(inputs.output_basename).$(inputs.tool_name).INFO_stripped.vcf.gz &&
       tabix $(inputs.output_basename).$(inputs.tool_name).INFO_stripped.vcf.gz) ||
       (echo "Check errors, likely does not have INFO, trying to pass input instead" >&2; cp $(inputs.input_vcf.path) .;

diff --git a/tools/gatk_variant_filter.cwl b/tools/gatk_variant_filter.cwl
@@ -6,8 +6,8 @@ requirements:
   - class: ShellCommandRequirement
   - class: InlineJavascriptRequirement
   - class: ResourceRequirement
-    ramMin: 8000
-    coresMin: 4
+    ramMin: 16000
+    coresMin: 8
   - class: DockerRequirement
     dockerPull: 'pgc-images.sbgenomics.com/d3b-bixu/gatk:4.1.1.0'
 

diff --git a/tools/generic_rename_outputs.cwl b/tools/generic_rename_outputs.cwl
@@ -6,8 +6,8 @@ requirements:
   - class: ShellCommandRequirement
   - class: InlineJavascriptRequirement
   - class: ResourceRequirement
-    ramMin: 2000
-    coresMin: 1
+    ramMin: 16000
+    coresMin: 8
 
 baseCommand: []
 arguments:

diff --git a/tools/hotspots_annotation.cwl b/tools/hotspots_annotation.cwl
@@ -49,7 +49,7 @@ inputs:
   csq_class: { type: 'string?', inputBinding: { position: 2, prefix: "--csq_class" }, doc: "Overrides the name of the CSQ field that stores variant class information" }
 
   # Resource Control
-  ram: { type: 'int?', default: 2, doc: "GB of RAM to allocate to this task." }
-  cores: { type: 'int?', default: 1, doc: "CPU cores to allocate to this task." }
+  ram: { type: 'int?', default: 16, doc: "GB of RAM to allocate to this task." }
+  cores: { type: 'int?', default: 8, doc: "CPU cores to allocate to this task." }
 outputs:
   hotspots_vcf: { type: 'File', outputBinding: { glob: '*.gz', outputEval: '$(inputs.disable_hotspot_annotation ? inputs.input_vcf : self)' }, secondaryFiles: [.tbi] }
diff --git a/tools/kf_mskcc_vcf2maf.cwl b/tools/kf_mskcc_vcf2maf.cwl
@@ -5,8 +5,8 @@ requirements:
   - class: ShellCommandRequirement
   - class: InlineJavascriptRequirement
   - class: ResourceRequirement
-    ramMin: 4000
-    coresMin: 2
+    ramMin: 16000
+    coresMin: 8
   - class: DockerRequirement
     dockerPull: 'pgc-images.sbgenomics.com/d3b-bixu/kf_vcf2maf:v1.0.3'
   - class: InitialWorkDirRequirement

diff --git a/tools/normalize_vcf.cwl b/tools/normalize_vcf.cwl
@@ -6,8 +6,8 @@ requirements:
   - class: ShellCommandRequirement
   - class: InlineJavascriptRequirement
   - class: ResourceRequirement
-    ramMin: 8000
-    coresMin: 4
+    ramMin: 16000
+    coresMin: 8
   - class: DockerRequirement
     dockerPull: 'pgc-images.sbgenomics.com/d3b-bixu/vcfutils:latest'