Merge pull request #14 from kids-first/bug/mb-fix-hard-filter

🐛 fix hard filter
kids-first · May 20, 2024 · f777ca0 · f777ca0
2 parents c380c66 + 9b27571
commit f777ca0
Show file tree

Hide file tree

Showing 6 changed files with 47 additions and 37 deletions.
diff --git a/docs/dockers_consensus.md b/docs/dockers_consensus.md
@@ -3,7 +3,7 @@
 TOOL|DOCKER
 -|-
 bcftools_annotate.cwl|pgc-images.sbgenomics.com/d3b-bixu/vcfutils:latest
-bcftools_filter_vcf.cwl|pgc-images.sbgenomics.com/d3b-bixu/bvcftools:latest
+bcftools_filter_vcf.cwl|pgc-images.sbgenomics.com/d3b-bixu/bcftools:1.20
 bcftools_strip_ann.cwl|pgc-images.sbgenomics.com/d3b-bixu/vcfutils:latest
 echtvar_anno.cwl|pgc-images.sbgenomics.com/d3b-bixu/echtvar:0.2.0
 generic_rename_outputs.cwl|None

diff --git a/docs/dockers_somatic.md b/docs/dockers_somatic.md
@@ -3,7 +3,7 @@
 TOOL|DOCKER
 -|-
 add_strelka2_fields.cwl|pgc-images.sbgenomics.com/d3b-bixu/add-strelka2-fields:1.0.0
-bcftools_filter_vcf.cwl|pgc-images.sbgenomics.com/d3b-bixu/bvcftools:latest
+bcftools_filter_vcf.cwl|pgc-images.sbgenomics.com/d3b-bixu/bcftools:1.20
 bcftools_strip_ann.cwl|pgc-images.sbgenomics.com/d3b-bixu/vcfutils:latest
 echtvar_anno.cwl|pgc-images.sbgenomics.com/d3b-bixu/echtvar:0.2.0
 gatk_variant_filter.cwl|pgc-images.sbgenomics.com/d3b-bixu/gatk:4.1.1.0

diff --git a/tools/bcftools_filter_vcf.cwl b/tools/bcftools_filter_vcf.cwl
@@ -5,46 +5,55 @@ doc: "More generic tool to take in an include expression and optionally an exclu
 requirements:
   - class: ShellCommandRequirement
   - class: DockerRequirement
-    dockerPull: 'pgc-images.sbgenomics.com/d3b-bixu/bvcftools:latest'
+    dockerPull: 'pgc-images.sbgenomics.com/d3b-bixu/bcftools:1.20'
   - class: ResourceRequirement
-    ramMin: 1000
-    coresMin: 1
+    ramMin: 16000
+    coresMin: 8
   - class: InlineJavascriptRequirement
-baseCommand: [bash -c 'set -eo pipefail &&]
+  - class: InitialWorkDirRequirement
+    listing:
+      - entryname: "run_filter.sh"
+        entry: |
+          #!/usr/bin/env bash
+          set -xeo pipefail
+
+          command=cat
+          if [[ $(inputs.sample_name) != null ]]
+          then
+            command="bcftools view --threads $(inputs.threads) -s $(inputs.sample_name)"
+          fi
+          $command $(inputs.input_vcf.path) \
+            | bcftools view \
+              --threads $(inputs.threads) \
+              -O $(inputs.output_type) \
+              -o $(inputs.output_basename).bcf_filtered.$(inputs.output_type == "v" ? "vcf" : inputs.output_type == "z" ? "vcf.gz" : inputs.output_type == "b" ? "bcf.gz" : "bcf") \
+              $(inputs.exclude_expression == null ? "" : "--exclude " + "'" + inputs.exclude_expression + "'") \
+              $(inputs.include_expression == null ? "" : "--include " + "'" + inputs.include_expression + "'") \
+              $(inputs.filter_expression == null ? "" : "-f " + inputs.filter_expression)
+          if [[ $(inputs.output_type) == z ]]
+          then
+            tabix $(inputs.output_basename).bcf_filtered.vcf.gz
+          fi
+
+baseCommand: []
 arguments:
-  - position: 1
+  - position: 0
     shellQuote: false
     valueFrom: >-
-      ${
-        var out_base = inputs.output_basename;
-        if (out_base == null){
-          out_base = inputs.input_vcf.nameroot + ".bcf_filtered"
-        }
-        var cmd = "bcftools view ";
-        if (inputs.include_expression != null){
-            cmd += "--include \"" + inputs.include_expression + "\" " + inputs.input_vcf.path;
-            if (inputs.exclude_expression != null){
-                cmd += " | bcftools view --exclude \"" + inputs.exclude_expression + "\"' -O z > " + out_base + ".vcf.gz;";
-            } else {
-                cmd += " -O z > " + out_base + ".vcf.gz;";
-            }
-        } else if (inputs.include_expression == null && inputs.exclude_expression != null){
-            cmd += "--exclude \"" + inputs.exclude_expression + "\" " + inputs.input_vcf.path + " -O z > " + out_base + ".vcf.gz;";
-        } else if (inputs.include_expression == null && inputs.exclude_expression == null){
-            cmd = "cp " + inputs.input_vcf.path + " ./" + out_base + ".vcf.gz;";
-        }
-        cmd += "tabix " + out_base + ".vcf.gz;'"
-        return cmd;
-      }
+      /bin/bash run_filter.sh
 
 inputs:
   input_vcf: File
-  include_expression: ['null', string]
-  exclude_expression: ['null', string]
-  output_basename: ['null', string]
+  include_expression: { type: 'string?', doc: "See bcftools docs for valid expression. Can't be used at the same time as exclude_expression. Use double quotes when a string needs to be quoted"}
+  threads: { type: 'int?', default: 4 }
+  exclude_expression: { type: 'string?', doc: "See bcftools docs for valid expression. Can't be used at the same time as include_expression.  Use double quotes when a string needs to be quoted"}
+  filter_expression: { type: 'string?', doc: "Add values from FILTER field to subset on"}
+  output_type: { type: [ 'null', {type: enum, name: output_type, symbols: [ "u", "b", "v", "z"]}]}
+  sample_name: { type: 'string?', doc: "csv string of samples if user wishes to apply filtering to and output specific samples"}
+  output_basename: string
 outputs:
   filtered_vcf:
     type: File
     outputBinding:
-      glob: '*.vcf.gz'
-    secondaryFiles: [.tbi]
+      glob: "*.{v,b}cf{,.gz}"
+    secondaryFiles: ['.tbi?']
diff --git a/tools/kf_mskcc_vcf2maf.cwl b/tools/kf_mskcc_vcf2maf.cwl
@@ -20,15 +20,16 @@ arguments:
   - position: 1
     shellQuote: false
     valueFrom: >-
-      $(inputs.input_vcf.path) > input_file.vcf
+      > input_file.vcf
       && perl vcf2maf.pl
       --input-vcf input_file.vcf
       --output-maf $(inputs.output_basename).$(inputs.tool_name).vep.maf
 
 inputs:
   reference: { type: 'File',  secondaryFiles: [.fai], doc: "Fasta genome assembly with index",
     inputBinding: {position: 2, prefix: "--ref-fasta"} }
-  input_vcf: { type: 'File', secondaryFiles: [.tbi], doc: "VEP annotated vcf file." }
+  input_vcf: { type: File, doc: "VEP annotated vcf file.", 
+    inputBinding: { position: 0 } }
   output_basename: string
   tumor_id: { type: string, inputBinding: {position: 3, prefix: "--tumor-id"} }
   normal_id: { type: string, inputBinding: {position: 4, prefix: "--normal-id"} }

diff --git a/workflows/kfdrc-germline-snv-annot-workflow.cwl b/workflows/kfdrc-germline-snv-annot-workflow.cwl
@@ -361,6 +361,6 @@ sbg:license: Apache License 2.0
 sbg:publisher: KFDRC
 
 "sbg:links":
-- id: 'https://github.com/kids-first/kids-first/kf-annotation-tools/releases/tag/v1.1.0'
+- id: 'https://github.com/kids-first/kids-first/kf-annotation-tools/releases/tag/v1.2.2'
   label: github-release
 
diff --git a/workflows/kfdrc-somatic-snv-annot-workflow.cwl b/workflows/kfdrc-somatic-snv-annot-workflow.cwl
@@ -352,5 +352,5 @@ $namespaces:
 "sbg:license": Apache License 2.0
 "sbg:publisher": KFDRC
 "sbg:links":
-- id: 'https://github.com/kids-first/kf-annotation-tools/releases/tag/v1.2.1'
+- id: 'https://github.com/kids-first/kf-annotation-tools/releases/tag/v1.2.2'
   label: github-release