Add GATK(3)/realignertargetcreator

modules/gatk/realignertargetcreator/main.nf

@@ -8,12 +8,12 @@ process GATK_REALIGNERTARGETCREATOR {
         'quay.io/biocontainers/gatk:3.5--hdfd78af_11' }"
 
     input:
-    tuple val(meta), path(bam)
+    tuple val(meta), path(bam), path(bai)
-    tuple val(meta), path(reference)
+    tuple path(fasta), path(fasta_fai), path(fasta_dict)
-    tuple val(meta), path(known_vcf)
+    path(known_vcf)
 
     output:
-    tuple val(meta), path("*.bam"), emit: bam
+    tuple val(meta), path("*.intervals"), emit: intervals
     path "versions.yml"                 , emit: versions
 
     when:
@@ -22,16 +22,16 @@ process GATK_REALIGNERTARGETCREATOR {
     script:
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
-    def known = known_vcf ? "-known ${known_vcf}" ? ""
+    def known = known_vcf ? "-known ${known_vcf}" : ""
     if ("$bam" == "${prefix}.bam") error "Input and output names are the same, set prefix in module configuration to disambiguate!"
 
     """
     gatk3 \\
-        -T RealigerTargetCreator \\
+        -T RealignerTargetCreator \\
-        -nt ${task.cpus}
+        -nt ${task.cpus} \\
         -I ${bam} \\
-        -R ${reference} \\
+        -R ${fasta} \\
-        -o ${prefix}.bam \\
+        -o ${prefix}.intervals \\
         ${known} \\
         $args
 

modules/gatk/realignertargetcreator/meta.yml

@@ -1,51 +1,64 @@
 name: "gatk_realignertargetcreator"
-## TODO nf-core: Add a description of the module and list keywords
+description: Generates a list of locations that should be considered for local realignment prior genotyping.
-description: write your description here
 keywords:
-  - sort
+  - bam
+  - vcf
+  - variant calling
+  - indel
+  - realignment
+  - targets
 tools:
   - "gatk":
-      ## TODO nf-core: Add a description and other details for the software below
       description: "The full Genome Analysis Toolkit (GATK) framework, license restricted."
-      homepage: "None"
+      homepage: "https://gatk.broadinstitute.org/hc/en-us"
-      documentation: "None"
+      documentation: "https://github.com/broadinstitute/gatk-docs"
-      tool_dev_url: "None"
+      licence: "['https://software.broadinstitute.org/gatk/download/licensing', 'BSD', 'https://www.broadinstitute.org/gatk/about/#licensing']"
-      doi: ""
-      licence: "['https://software.broadinstitute.org/gatk/download/licensing', 'https://www.broadinstitute.org/gatk/about/#licensing', 'BSD']"
 
-## TODO nf-core: Add a description of all of the variables used as input
 input:
-  # Only when we have meta
   - meta:
       type: map
       description: |
         Groovy Map containing sample information
         e.g. [ id:'test', single_end:false ]
-  # 
-  ## TODO nf-core: Delete / customise this example input
   - bam:
       type: file
-      description: BAM/CRAM/SAM file
+      description: Sorted and indexed BAM/CRAM/SAM file
-      pattern: "*.{bam,cram,sam}"
+      pattern: "*.bam"
+  - bai:
+      type: file
+      description: BAM index file
+      pattern: "*.bai"
+  - fasta:
+      type: file
+      description: Reference file used to generate BAM file
+      pattern: ".{fasta,fa,fna}"
+  - fasta_fai:
+      type: file
+      description: Index of reference file used to generate BAM file
+      pattern: ".fai"
+  - dict:
+      type: file
+      description: GATK dict file for reference
+      pattern: ".dict"
+  - known_vcf:
+      type: file
+      description: Optional input VCF file(s) with known indels
+      pattern: ".vcf"
 
-## TODO nf-core: Add a description of all of the variables used as output
 output:
-  #Only when we have meta
   - meta:
       type: map
       description: |
         Groovy Map containing sample information
         e.g. [ id:'test', single_end:false ]
-  # 
   - versions:
       type: file
       description: File containing software versions
       pattern: "versions.yml"
-  ## TODO nf-core: Delete / customise this example output
+  - intervals:
-  - bam:
       type: file
-      description: Sorted BAM/CRAM/SAM file
+      description: File containg intervals that represent sites of extant and potential indels.
-      pattern: "*.{bam,cram,sam}"
+      pattern: "*.intervals"
 
 authors:
   - "@jfy133"

tests/modules/gatk/realignertargetcreator/main.nf

@@ -8,8 +8,14 @@ workflow test_gatk_realignertargetcreator {
 
     input = [
         [ id:'test', single_end:false ], // meta map
-        file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true)
+        file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true),
+        file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam_bai'], checkIfExists: true)
+    ]
+    reference = [
+        file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true),
+        file(params.test_data['sarscov2']['genome']['genome_fasta_fai'], checkIfExists: true),
+        file(params.test_data['sarscov2']['genome']['genome_dict'], checkIfExists: true),
     ]
 
-    GATK_REALIGNERTARGETCREATOR ( input )
+    GATK_REALIGNERTARGETCREATOR ( input, reference, [] )
 }

tests/modules/gatk/realignertargetcreator/test.yml

@@ -1,14 +1,8 @@
-## TODO nf-core: Please run the following command to build this file:
+- name: gatk realignertargetcreator test_gatk_realignertargetcreator
-#                nf-core modules create-test-yml gatk/realignertargetcreator
-- name: "gatk realignertargetcreator"
   command: nextflow run ./tests/modules/gatk/realignertargetcreator -entry test_gatk_realignertargetcreator -c ./tests/config/nextflow.config  -c ./tests/modules/gatk/realignertargetcreator/nextflow.config
   tags:
-    - "gatk"
+    - gatk
-    #
+    - gatk/realignertargetcreator
-    - "gatk/realignertargetcreator"
-    #
   files:
-    - path: "output/gatk/test.bam"
+    - path: output/gatk/test.intervals
-      md5sum: e667c7caad0bc4b7ac383fd023c654fc
+      md5sum: 7aa7a1b235a510e6591e262382086bf8
-    - path: output/gatk/versions.yml
-      md5sum: a01fe51bc4c6a3a6226fbf77b2c7cf3b