add somatic SV calling to somatic workflow

docs/index.md

@@ -12,7 +12,7 @@ MultiQC), adapter clipping (using cutadapt), mapping (using BWA mem or
 bwakit) and variant calling (based on the
 [GATK Best Practice](https://software.broadinstitute.org/gatk/best-practices/)
 for germline calling, and using a variety of callers for somatic calling).
-Optionally, the somatic workflow can also perform CNV calling.
+Optionally, the somatic workflow can also perform CNV calling and SV calling.
 
 This workflow is part of [BioWDL](https://biowdl.github.io/)
 developed by the SASC team
@@ -316,6 +316,9 @@ This workflow will produce a number of directories and files:
     It contains the vcf file for the sample if germline.wdl was used and
     a single sample vcf was produced.
     It also contains a directory per readgroup.
+    - **structural-variants**: Structural variant calling results per
+      caller and the merged results from SURVIVOR. Only present if 
+      `germline.wdl` is used with SV calling enabled.
     - **CNVcalling**: Contains the CNV calling results for this sample
       and its control sample. Only present if `somatic.wdl` is used with
       CNV calling enabled.
@@ -329,6 +332,11 @@ This workflow will produce a number of directories and files:
 - **PON**: A generated panel of normals and the preprocessed intervals.
   Only present if `somatic.wdl` is used with CNV calling enabled and no
   PON or preprocessed intervals were provided in the inputs.
+- **somatic-sv-calling**: Somatic SV calling results. Only present if
+  `somatic.wdl` is used with SV calling enabled.
+  - **<sample>**: SV results for Delly and Manta per tumor sample.
+  - **gridss**: GRIDSS results per normal and the generated PON files 
+    (if no PON was provided in the inputs).
 
 ## Scattering
 This workflow performs scattering to speed up analysis on grid computing

germline.wdl

@@ -309,8 +309,10 @@ workflow Germline {
         cleverVcfs: {description: ""}
         matecleverVcfs: {description: ""}
         mantaVcfs: {description: ""}
+        smooveVcfs: {description: ""}
         dellyVcfs: {description: ""}
         survivorVcfs: {description: ""}
+        gridssVcfs: {description: ""}
         gridssVcfIndexes: {description: ""}
         SVunionVcfs: {description: ""}
         SVisecVcfs: {description: ""}

somatic.wdl

@@ -25,6 +25,7 @@ import "gatk-CNVcalling/CNV-PON.wdl" as cnvPon
 import "sample.wdl" as sampleWorkflow
 import "somatic-variantcalling/somatic-variantcalling.wdl" as somaticVariantcallingWorkflow
 import "structs.wdl" as structs
+import "structural-variantcalling/somatic.wdl" as somaticSvCalling
 import "tasks/biowdl.wdl" as biowdl
 import "tasks/bwa.wdl" as bwa
 import "tasks/common.wdl" as common
@@ -45,6 +46,7 @@ workflow Somatic {
         Boolean umiDeduplication = false
         Boolean collectUmiStats = false
         Boolean performCnvCalling = false
+        Boolean performSvCalling = false
         String platform = "illumina"
         Boolean useBwaKit = false
         Boolean runStrelka = true
@@ -194,7 +196,17 @@ workflow Somatic {
                     sample = select_first([sample.control])
             }
 
+            # Collect the inputs for SV calling, so only paired samples are included.
+            String tumorIdsForSvCalling = sample.id
+            File tumorBamsForSvCalling = sampleWorkflow.recalibratedBam[casePosition.position]
+            File tumorBamIndexesForSvCalling = sampleWorkflow.recalibratedBamIndex[casePosition.position]
+            String controlIdsForSvCalling = select_first([sample.control])
+            File controlBamsForSvCalling = sampleWorkflow.recalibratedBam[controlPostition.position]
+            File controlBamIndexesForSvCalling = sampleWorkflow.recalibratedBamIndex[controlPostition.position]
+            Pair[String, String] tumorControlPairsForSvCalling = (sample.id, select_first([sample.control]))
         }
+
+        # Allow SNV calling on tumor-only samples as well.
         Int controlPos = select_first([controlPostition.position, 0])
         String? controlSample = if (defined(sample.control)) then sampleIds[controlPos] else DONOTDEFINETHISSTRING
         File? controlBam = if (defined(sample.control)) then sampleWorkflow.recalibratedBam[controlPos] else DONOTDEFINETHISFILE
@@ -245,6 +257,23 @@ workflow Somatic {
         }
     }
 
+    if (performSvCalling) {
+        call somaticSvCalling.SomaticSvCalling as SVs {
+            input:
+                normalIds = select_all(controlIdsForSvCalling),
+                normalBams = select_all(controlBamsForSvCalling),
+                normalBamIndexes = select_all(controlBamIndexesForSvCalling),
+                tumorIds = select_all(tumorIdsForSvCalling),
+                tumorBams = select_all(tumorBamsForSvCalling),
+                tumorBamIndexes = select_all(tumorBamIndexesForSvCalling),
+                pairs = select_all(tumorControlPairsForSvCalling),
+                referenceFasta = refFasta,
+                referenceFastaFai = refFastaFai,
+                bwaIndex = bwidx,
+                outputDir = outputDir + "/somatic-sv-calling"
+        }
+    }
+
     call multiqc.MultiQC as multiqcTask {
         input:
             reports = flatten(sampleWorkflow.reports),
@@ -322,6 +351,7 @@ workflow Somatic {
         dbsnpVCF: {description: "dbsnp VCF file used for checking known sites.", category: "required"}
         dbsnpVCFIndex: {description: "Index (.tbi) file for the dbsnp VCF. Will be created automatically if not present.", category: "common"}
         performCnvCalling: {description: "Whether or not CNV calling should be performed.", category: "common"}
+        performSvCalling: {description: "Whether or not SV calling should be performed.", category: "common"}
         platform: {description: "The platform used for sequencing.", category: "advanced"}
         useBwaKit: {description: "Whether or not BWA kit should be used. If false BWA mem will be used.", category: "advanced"}
         runStrelka: {description: "Whether or not to run Strelka.", category: "common"}