Add GATK UnifiedGenotyper

modules/gatk/unifiedgenotyper/main.nf

@@ -12,15 +12,14 @@ process GATK_UNIFIEDGENOTYPER {
     path(fasta)
     path(fai)
     path(dict)
-    path(known_vcf)
     path(intervals)
     path(contamination)
-    path(dbsnps)
+    path(dbsnp)
     path(comp)
 
     output:
     tuple val(meta), path("*.vcf.gz"), emit: vcf
-    path "versions.yml"           , emit: versions
+    path "versions.yml"              , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
@@ -29,7 +28,7 @@ process GATK_UNIFIEDGENOTYPER {
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
     def contamination_file = contamination ? "-contaminationFile ${contamination}" : ""
-    def dbsnps_file = dbsnps ? "--dbsnp ${dbsnps}" : ""
+    def dbsnp_file = dbsnp ? "--dbsnp ${dbsnp}" : ""
     def comp_file = comp ? "--comp ${comp}" : ""
     def intervals_file = intervals ? "--intervals ${intervals}" : ""
 
@@ -48,9 +47,9 @@ process GATK_UNIFIEDGENOTYPER {
         -I ${input} \\
         -R ${fasta} \\
         ${contamination_file} \\
-        ${dbsnps_file} \\
+        ${dbsnp_file} \\
         ${comp_file} \\
-        ${intervals_file}
+        ${intervals_file} \\
         -o ${prefix}.vcf \\
         $args
 
@@ -58,7 +57,7 @@ process GATK_UNIFIEDGENOTYPER {
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
-        gatk: \$(echo \$(samtools --version 2>&1) | sed 's/^.*samtools //; s/Using.*\$//' ))
+        gatk: \$(echo \$(gatk3 --version))
     END_VERSIONS
     """
 }

modules/gatk/unifiedgenotyper/meta.yml

@@ -3,9 +3,6 @@ keywords:
   - bam
   - vcf
   - variant calling
-  - indel
-  - realignment
-  - targets
 tools:
   - "gatk":
       description: "The full Genome Analysis Toolkit (GATK) framework, license restricted."
@@ -39,10 +36,22 @@ input:
       type: file
       description: GATK dict file for reference
       pattern: ".dict"
-  - known_vcf:
+  - intervals:
       type: file
-      description: Optional input VCF file(s) with known indels
+      description: Bed file with the genomic regions included in the library (optional)
-      pattern: ".vcf"
+      pattern: "*.intervals"
+  - contamination:
+      type: file
+      description: Tab-separated file containing fraction of contamination in sequencing data (per sample) to aggressively remove
+      pattern: "*"
+  - dbsnps:
+      type: file
+      description: VCF file containing known sites (optional)
+      pattern: "*"
+  - comp:
+      type: file
+      description: Comparison VCF file (optional)
+      pattern: "*"
 
 output:
   - meta:

tests/modules/gatk/unifiedgenotyper/main.nf

@@ -5,11 +5,14 @@ nextflow.enable.dsl = 2
 include { GATK_UNIFIEDGENOTYPER } from '../../../../modules/gatk/unifiedgenotyper/main.nf'
 
 workflow test_gatk_unifiedgenotyper {
-    
-    input = [
-        [ id:'test', single_end:false ], // meta map
-        file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true)
-    ]
 
-    GATK_UNIFIEDGENOTYPER ( input )
+    input     = [ [ id:'test' ], // meta map
+                  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),
+                ]
+    fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    fai = file(params.test_data['sarscov2']['genome']['genome_fasta_fai'], checkIfExists: true)
+    dict = file(params.test_data['sarscov2']['genome']['genome_dict'], checkIfExists: true)
+
+    GATK_UNIFIEDGENOTYPER ( input, fasta, fai, dict, [], [], [], [])
 }

tests/modules/gatk/unifiedgenotyper/test.yml

@@ -1,14 +1,9 @@
-## TODO nf-core: Please run the following command to build this file:
+- name: gatk unifiedgenotyper test_gatk_unifiedgenotyper
-#                nf-core modules create-test-yml gatk/unifiedgenotyper
+  command: nextflow run ./tests/modules/gatk/unifiedgenotyper -entry test_gatk_unifiedgenotyper -c ./tests/config/nextflow.config  -c ./tests/modules/gatk/unifiedgenotyper/nextflow.config
-- name: "gatk unifiedgenotyper"
-  command: nextflow run ./tests/modules/gatk/unifiedgenotyper -entry test_gatk_unifiedgenotyper -c ./tests/config/nextflow.config -c ./tests/modules/gatk/unifiedgenotyper/nextflow.config
   tags:
-    - "gatk"
+    - gatk
-    #
+    - gatk/unifiedgenotyper
-    - "gatk/unifiedgenotyper"
-    #
   files:
-    - path: "output/gatk/test.bam"
+    - path: output/gatk/test.vcf.gz
-      md5sum: e667c7caad0bc4b7ac383fd023c654fc
+      contains:
-    - path: output/gatk/versions.yml
+        - "#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	test"
-      md5sum: a01fe51bc4c6a3a6226fbf77b2c7cf3b