TSPS-326 add wdl to mask or subset a vcf given a bed file (#136)

Co-authored-by: Jose Soto <jsoto@broadinstitute.org>

.dockstore.yml

@@ -35,3 +35,10 @@ workflows:
     authors:
       - name: Terra Scientific Services
         email: teaspoons-developers@broadinstitute.org
+
+  - subclass: WDL
+    name: SubsetVcfByBedFile
+    primaryDescriptorPath: /pipelines/imputation/simulatedData/SubsetVcfByBedFile.wdl
+    authors:
+      - name: Terra Scientific Services
+        email: teaspoons-developers@broadinstitute.org

pipelines/imputation/scientificValidation/README.md

@@ -13,3 +13,23 @@ This wdl is basically a wrapper around that tool/image
 #### Outputs
 * recombined_reference_panel - output vcf after mitigation 
 algorithm has been run
+
+
+## SubsetVcfByBedFile
+### Purpose
+This wdl is meant to be used to subset a vcf down
+to sites provided through a bed file.  This wdl does
+not interact with headers or annotations mostly because
+the only really "required" header is the dictionary
+and that gets transferred across and the imputation
+tool only look at GT and no info/format fields so
+we can just leave them be and have it not affect
+
+#### Inputs
+* input_vcf - input file to be subset
+* input_vcf_index 
+* bed_file - bed file containing intervals to subset by
+
+#### Outputs
+* subset_vcf - subsetted vcf
+* subset_vcf_index 

pipelines/imputation/scientificValidation/ReshapeReferencePanel.wdl

@@ -31,7 +31,7 @@ task ReshapeReferencePanel {
 
         Int disk_size_gb = ceil(3*size(ref_panel_vcf, "GiB")) + 20
         Int cpu = 1
-        Int memory_mb = 8000
+        Int memory_mb = 6000
     }
 
     command {

pipelines/imputation/scientificValidation/SubsetVcfByBedFile.wdl

@@ -0,0 +1,61 @@
+version 1.0
+
+# This script is under review. It is not actively tested or maintained at this time.
+workflow SubsetVcfByBedFile {
+    input {
+        File input_vcf
+        File input_vcf_index
+        File bed_file
+    }
+
+    call BcftoolsSubsetVcf {
+        input:
+            input_vcf = input_vcf,
+            input_vcf_index = input_vcf_index,
+            bed_file = bed_file
+    }
+
+    output {
+        File subset_vcf = BcftoolsSubsetVcf.output_vcf
+        File subset_vcf_index = BcftoolsSubsetVcf.output_vcf_index
+    }
+}
+
+task BcftoolsSubsetVcf {
+    input {
+        File input_vcf
+        File input_vcf_index
+        File bed_file
+
+        Int disk_size_gb = ceil(3 * size(input_vcf, "GiB")) + 20
+        Int cpu = 1
+        Int memory_mb = 6000
+    }
+
+    String basename = basename(input_vcf, '.vcf.gz')
+
+    command {
+        set -e -o pipefail
+
+        bcftools view \
+        -R ~{bed_file} \
+        -O z \
+        -o ~{basename}.subset.vcf.gz \
+        ~{input_vcf}
+
+        bcftools index -t ~{basename}.subset.vcf.gz
+
+    }
+
+    output {
+        File output_vcf = "~{basename}.subset.vcf.gz"
+        File output_vcf_index = "~{basename}.subset.vcf.gz.tbi"
+    }
+
+    runtime {
+        docker: "us.gcr.io/broad-gatk/gatk:4.5.0.0"
+        disks: "local-disk ${disk_size_gb} HDD"
+        memory: "${memory_mb} MiB"
+        cpu: cpu
+    }
+}