From 989fb7d5b7db55d1fa305fad23877f95591f1e1a Mon Sep 17 00:00:00 2001
From: suzannejin <43495634+suzannejin@users.noreply.github.com>
Date: Fri, 26 Mar 2021 12:46:46 +0100
Subject: [PATCH] modify test path - module gatk4   +  added new module
 gatk4/haplotypecaller (#382)

* new module: samtools/fastq

* solve conflict: pytest_software.yml

* solve linting conflicts

* solved EditorConfig linting problem

* Module samtools/fastq:
	* output compressed fastq.gz file(s)
	* add if conditionals for single/paired reads

* samtools/fastq: modified test.yml

* samtools/fastq: modified main.nf to avoid duplicated part of the script section

* fix README.md

* modify test path gatk4

* fix config

* Add new module gatk4/haplotypecaller

* solve check issues

* fix test.yml file

* fix test.yml gz.tbi

* Update software/gatk4/haplotypecaller/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gatk4/haplotypecaller/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gatk4/haplotypecaller/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gatk4/haplotypecaller/meta.yml

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gatk4/haplotypecaller/meta.yml

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update tests/software/gatk4/haplotypecaller/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Apply suggestions from code review

Co-authored-by: suzannejin <suzanne.jin@crg.eu>
Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>
---
 software/gatk4/haplotypecaller/functions.nf   | 60 ++++++++++++++++++
 software/gatk4/haplotypecaller/main.nf        | 52 +++++++++++++++
 software/gatk4/haplotypecaller/meta.yml       | 63 +++++++++++++++++++
 tests/config/pytest_software.yml              |  4 ++
 tests/software/gatk4/applybqsr/main.nf        | 26 ++++----
 tests/software/gatk4/applybqsr/test.yml       | 12 ++--
 tests/software/gatk4/baserecalibrator/main.nf | 62 +++++++++---------
 .../software/gatk4/baserecalibrator/test.yml  |  8 +--
 .../software/gatk4/bedtointervallist/main.nf  |  4 +-
 .../software/gatk4/bedtointervallist/test.yml |  4 +-
 .../gatk4/createsequencedictionary/main.nf    |  2 +-
 .../gatk4/createsequencedictionary/test.yml   |  8 +--
 tests/software/gatk4/fastqtosam/main.nf       |  6 +-
 tests/software/gatk4/fastqtosam/test.yml      |  4 +-
 tests/software/gatk4/haplotypecaller/main.nf  | 17 +++++
 tests/software/gatk4/haplotypecaller/test.yml | 13 ++++
 .../software/gatk4/mergebamalignment/main.nf  |  8 +--
 .../software/gatk4/mergebamalignment/test.yml |  6 +-
 tests/software/gatk4/mergevcfs/main.nf        | 14 ++---
 tests/software/gatk4/mergevcfs/test.yml       | 13 ++--
 tests/software/gatk4/revertsam/main.nf        |  2 +-
 tests/software/gatk4/revertsam/test.yml       |  6 +-
 tests/software/gatk4/samtofastq/main.nf       |  4 +-
 tests/software/gatk4/samtofastq/test.yml      | 16 ++---
 tests/software/gatk4/splitncigarreads/main.nf |  8 +--
 .../software/gatk4/splitncigarreads/test.yml  |  6 +-
 26 files changed, 319 insertions(+), 109 deletions(-)
 create mode 100644 software/gatk4/haplotypecaller/functions.nf
 create mode 100644 software/gatk4/haplotypecaller/main.nf
 create mode 100644 software/gatk4/haplotypecaller/meta.yml
 create mode 100644 tests/software/gatk4/haplotypecaller/main.nf
 create mode 100644 tests/software/gatk4/haplotypecaller/test.yml

diff --git a/software/gatk4/haplotypecaller/functions.nf b/software/gatk4/haplotypecaller/functions.nf
new file mode 100644
index 00000000..f177f0c8
--- /dev/null
+++ b/software/gatk4/haplotypecaller/functions.nf
@@ -0,0 +1,60 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.args3         = args.args3 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}
diff --git a/software/gatk4/haplotypecaller/main.nf b/software/gatk4/haplotypecaller/main.nf
new file mode 100644
index 00000000..d8938297
--- /dev/null
+++ b/software/gatk4/haplotypecaller/main.nf
@@ -0,0 +1,52 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process GATK4_HAPLOTYPECALLER {
+    tag "$meta.id"
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:meta.id) }
+
+    conda (params.enable_conda ? "bioconda::gatk4=4.2.0.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/gatk4:4.2.0.0--0"
+    } else {
+        container "quay.io/biocontainers/gatk4:4.2.0.0--0"
+    }
+
+    input:
+    tuple val(meta), path(bam), path(bai)
+    path fasta
+    path fai
+    path dict
+
+    output:
+    tuple val(meta), path("*.vcf.gz"), emit: vcf
+    tuple val(meta), path("*.tbi")   , emit: tbi
+    path "*.version.txt"             , emit: version
+
+    script:
+    def software  = getSoftwareName(task.process)
+    def prefix    = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def avail_mem = 3
+    if (!task.memory) {
+        log.info '[GATK HaplotypeCaller] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = task.memory.giga
+    }
+    """
+    gatk \\
+        --java-options "-Xmx${avail_mem}g" \\
+        HaplotypeCaller \\
+        -R $fasta \\
+        -I $bam \\
+        -O ${prefix}.vcf.gz \\
+        $options.args
+
+    gatk --version | grep Picard | sed "s/Picard Version: //g" > ${software}.version.txt
+    """
+}
diff --git a/software/gatk4/haplotypecaller/meta.yml b/software/gatk4/haplotypecaller/meta.yml
new file mode 100644
index 00000000..f0fc3910
--- /dev/null
+++ b/software/gatk4/haplotypecaller/meta.yml
@@ -0,0 +1,63 @@
+name: gatk4_haplotypecaller
+description: Call germline SNPs and indels via local re-assembly of haplotypes
+keywords:
+  - gatk4
+  - haplotypecaller
+  - haplotype
+tools:
+  - gatk4:
+    description: |
+        Developed in the Data Sciences Platform at the Broad Institute, the toolkit offers a wide variety of tools
+        with a primary focus on variant discovery and genotyping. Its powerful processing engine
+        and high-performance computing features make it capable of taking on projects of any size.
+    homepage: https://gatk.broadinstitute.org/hc/en-us
+    documentation: https://gatk.broadinstitute.org/hc/en-us/categories/360002369672s
+    doi: 10.1158/1538-7445.AM2017-3590
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - bam:
+      type: file
+      description: BAM file
+      pattern: "*.bam"
+  - bai:
+      type: file
+      description: Index of BAM file
+      pattern: "*.bam.bai"
+  - fasta:
+      type: file
+      description: The reference fasta file
+      pattern: "*.fasta"
+  - fai:
+      type: file
+      description: Index of reference fasta file
+      pattern: "fasta.fai"
+  - dict:
+      type: file
+      description: GATK sequence dictionary
+      pattern: "*.dict"
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - vcf:
+      type: file
+      description: Compressed VCF file
+      pattern: "*.vcf.gz"
+  - tbi:
+      type: file
+      description: Index of VCF file
+      pattern: "*.vcf.gz.tbi"
+
+authors:
+  - "@suzannejin"
diff --git a/tests/config/pytest_software.yml b/tests/config/pytest_software.yml
index 554d30d8..16aa1a31 100755
--- a/tests/config/pytest_software.yml
+++ b/tests/config/pytest_software.yml
@@ -202,6 +202,10 @@ gatk4_fastqtosam:
   - software/gatk4/fastqtosam/**
   - tests/software/gatk4/fastqtosam/**
 
+gatk4_haplotypecaller:
+  - software/gatk4/haplotypecaller/**
+  - tests/software/gatk4/haplotypecaller/**
+
 gatk4_mergebamalignment:
   - software/gatk4/mergebamalignment/**
   - tests/software/gatk4/mergebamalignment/**
diff --git a/tests/software/gatk4/applybqsr/main.nf b/tests/software/gatk4/applybqsr/main.nf
index 0029a508..2d3e532c 100644
--- a/tests/software/gatk4/applybqsr/main.nf
+++ b/tests/software/gatk4/applybqsr/main.nf
@@ -6,25 +6,25 @@ include { GATK4_APPLYBQSR } from '../../../../software/gatk4/applybqsr/main.nf'
 
 workflow test_gatk4_applybqsr {
     input = [ [ id:'test' ], // meta map
-              file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.sorted.bam", checkIfExists: true),
-              file("${launchDir}/tests/data/genomics/sarscov2/illumina/gatk/test.table", checkIfExists: true) 
+              file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true),
+              file(params.test_data['sarscov2']['illumina']['test_baserecalibrator_table'], checkIfExists: true)
             ]
-    fasta = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
-    fai   = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta.fai", checkIfExists: true)
-    dict  = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.dict", 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)
 
     GATK4_APPLYBQSR ( input, fasta, fai, dict, [] )
 }
 
 workflow test_gatk4_applybqsr_intervals {
-  input     = [ [ id:'test' ], // meta map
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.sorted.bam", checkIfExists: true),
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/gatk/test.table", checkIfExists: true)  
-            ]
-  fasta     = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
-  fai       = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta.fai", checkIfExists: true)
-  dict      = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.dict", checkIfExists: true)
-  intervals = file("${launchDir}/tests/data/genomics/sarscov2/genome/bed/test.bed", checkIfExists: true)
+    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_baserecalibrator_table'], 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)
+    intervals = file(params.test_data['sarscov2']['genome']['test_bed'], checkIfExists: true)
 
   GATK4_APPLYBQSR ( input, fasta, fai, dict, intervals )
 }
diff --git a/tests/software/gatk4/applybqsr/test.yml b/tests/software/gatk4/applybqsr/test.yml
index 9744aaa0..b17ffbe2 100644
--- a/tests/software/gatk4/applybqsr/test.yml
+++ b/tests/software/gatk4/applybqsr/test.yml
@@ -1,18 +1,18 @@
 - name: gatk4 applybqsr test_gatk4_applybqsr
   command: nextflow run tests/software/gatk4/applybqsr -entry test_gatk4_applybqsr -c tests/config/nextflow.config
   tags:
-    - gatk4
     - gatk4_applybqsr
+    - gatk4
   files:
     - path: output/gatk4/test.bam
-      md5sum: 943a43de5f575a0435fe13072cbc15c8
+      md5sum: 139a5ffc7601ce3cfa7a896cd381542a
 
-- name: "\e[3~"
+- name: gatk4 applybqsr test_gatk4_applybqsr_intervals
   command: nextflow run tests/software/gatk4/applybqsr -entry test_gatk4_applybqsr_intervals -c tests/config/nextflow.config
   tags:
-    - gatk4
-    - gatk4_applybqsr_intervals
     - gatk4_applybqsr
+    - gatk4_applybqsr_intervals
+    - gatk4
   files:
     - path: output/gatk4/test.bam
-      md5sum: a92238a9a78da48bf437fd5b7d9790a3
+      md5sum: 5b1f6fa2525124c281f71e5a76d28482
diff --git a/tests/software/gatk4/baserecalibrator/main.nf b/tests/software/gatk4/baserecalibrator/main.nf
index c93ef044..302ddb1f 100644
--- a/tests/software/gatk4/baserecalibrator/main.nf
+++ b/tests/software/gatk4/baserecalibrator/main.nf
@@ -5,45 +5,45 @@ nextflow.enable.dsl = 2
 include { GATK4_BASERECALIBRATOR } from '../../../../software/gatk4/baserecalibrator/main.nf' addParams( options: [:] )
 
 workflow test_gatk4_baserecalibrator {
-  input     = [ [ id:'test' ], // meta map
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.sorted.bam", checkIfExists: true) 
-              ]
-  fasta     = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
-  fai       = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta.fai", checkIfExists: true)
-  dict      = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.dict", checkIfExists: true)
-  sites     = file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf.gz", checkIfExists: true)
-  sites_tbi = file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf.gz.tbi", checkIfExists: true)
+    input     = [ [ id:'test' ], // meta map
+                    file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam'], 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)
+    sites = file(params.test_data['sarscov2']['illumina']['test_vcf_gz'], checkIfExists: true)
+    sites_tbi = file(params.test_data['sarscov2']['illumina']['test_vcf_gz_tbi'], checkIfExists: true)
 
-  GATK4_BASERECALIBRATOR ( input, fasta, fai, dict, [], sites, sites_tbi )
+    GATK4_BASERECALIBRATOR ( input, fasta, fai, dict, [], sites, sites_tbi )
 }
 
 workflow test_gatk4_baserecalibrator_intervals {
-  input =   [ [ id:'test' ], // meta map
-              file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.sorted.bam", checkIfExists: true) 
-            ]
-  fasta     = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
-  fai       = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta.fai", checkIfExists: true)
-  dict      = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.dict", checkIfExists: true)
-  intervals = file("${launchDir}/tests/data/genomics/sarscov2/genome/bed/test.bed", checkIfExists: true)
-  sites     = file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf.gz", checkIfExists: true)
-  sites_tbi = file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf.gz.tbi", checkIfExists: true)
+    input     = [ [ id:'test' ], // meta map
+                  file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam'], 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)
+    intervals = file(params.test_data['sarscov2']['genome']['test_bed'], checkIfExists: true)
+    sites = file(params.test_data['sarscov2']['illumina']['test_vcf_gz'], checkIfExists: true)
+    sites_tbi = file(params.test_data['sarscov2']['illumina']['test_vcf_gz_tbi'], checkIfExists: true)
 
-  GATK4_BASERECALIBRATOR ( input, fasta, fai, dict, intervals, sites, sites_tbi )
+    GATK4_BASERECALIBRATOR ( input, fasta, fai, dict, intervals, sites, sites_tbi )
 }
 
 workflow test_gatk4_baserecalibrator_multiple_sites {
-  input     = [ [ id:'test' ], // meta map
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.sorted.bam", checkIfExists: true) 
-              ]
-  fasta     = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
-  fai       = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta.fai", checkIfExists: true)
-  dict      = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.dict", checkIfExists: true)
-  sites     = [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf.gz", checkIfExists: true),
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test2.vcf.gz", checkIfExists: true)
-              ]
-  sites_tbi = [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf.gz.tbi", checkIfExists: true),
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test2.vcf.gz.tbi", checkIfExists: true)
-              ]
+    input     = [ [ id:'test' ], // meta map
+                  file(params.test_data['sarscov2']['illumina']['test_paired_end_sorted_bam'], 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)
+    sites =     [ file(params.test_data['sarscov2']['illumina']['test_vcf_gz'], checkIfExists: true),
+                  file(params.test_data['sarscov2']['illumina']['test2_vcf_gz'], checkIfExists: true)
+                ]
+    sites_tbi = [ file(params.test_data['sarscov2']['illumina']['test_vcf_gz_tbi'], checkIfExists: true),
+                  file(params.test_data['sarscov2']['illumina']['test2_vcf_gz_tbi'], checkIfExists: true)
+                ]
 
   GATK4_BASERECALIBRATOR ( input, fasta, fai, dict, [], sites, sites_tbi )
 }
diff --git a/tests/software/gatk4/baserecalibrator/test.yml b/tests/software/gatk4/baserecalibrator/test.yml
index 61dd2283..c0011152 100644
--- a/tests/software/gatk4/baserecalibrator/test.yml
+++ b/tests/software/gatk4/baserecalibrator/test.yml
@@ -1,8 +1,8 @@
 - name: gatk4 baserecalibrator test_gatk4_baserecalibrator
   command: nextflow run tests/software/gatk4/baserecalibrator -entry test_gatk4_baserecalibrator -c tests/config/nextflow.config
   tags:
-    - gatk4
     - gatk4_baserecalibrator
+    - gatk4
   files:
     - path: output/gatk4/test.table
       md5sum: e2e43abdc0c943c1a54dae816d0b9ea7
@@ -10,9 +10,9 @@
 - name: gatk4 baserecalibrator test_gatk4_baserecalibrator_intervals
   command: nextflow run tests/software/gatk4/baserecalibrator -entry test_gatk4_baserecalibrator_intervals -c tests/config/nextflow.config
   tags:
-    - gatk4
-    - gatk4_baserecalibrator
     - gatk4_baserecalibrator_intervals
+    - gatk4_baserecalibrator
+    - gatk4
   files:
     - path: output/gatk4/test.table
       md5sum: 9ecb5f00a2229291705addc09c0ec231
@@ -21,8 +21,8 @@
   command: nextflow run tests/software/gatk4/baserecalibrator -entry test_gatk4_baserecalibrator_multiple_sites -c tests/config/nextflow.config
   tags:
     - gatk4_baserecalibrator_multiple_sites
-    - gatk4
     - gatk4_baserecalibrator
+    - gatk4
   files:
     - path: output/gatk4/test.table
       md5sum: e2e43abdc0c943c1a54dae816d0b9ea7
diff --git a/tests/software/gatk4/bedtointervallist/main.nf b/tests/software/gatk4/bedtointervallist/main.nf
index 4b8c32d9..37cb164d 100644
--- a/tests/software/gatk4/bedtointervallist/main.nf
+++ b/tests/software/gatk4/bedtointervallist/main.nf
@@ -6,9 +6,9 @@ include { GATK4_BEDTOINTERVALLIST } from '../../../../software/gatk4/bedtointerv
 
 workflow test_gatk4_bedtointervallist {
     input = [ [ id:'test' ], // meta map
-              [ file("${launchDir}/tests/data/genomics/sarscov2/genome/bed/test.bed", checkIfExists: true)] 
+              [ file(params.test_data['sarscov2']['genome']['test_bed'], checkIfExists: true) ]
             ]
-    dict  = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.dict", checkIfExists: true)
+    dict = file(params.test_data['sarscov2']['genome']['genome_dict'], checkIfExists: true)
 
     GATK4_BEDTOINTERVALLIST ( input, dict )
 }
diff --git a/tests/software/gatk4/bedtointervallist/test.yml b/tests/software/gatk4/bedtointervallist/test.yml
index 6612cd19..a8bb7bf0 100644
--- a/tests/software/gatk4/bedtointervallist/test.yml
+++ b/tests/software/gatk4/bedtointervallist/test.yml
@@ -1,5 +1,5 @@
-- name: gatk4 bedtointervallist
-  command: nextflow run ./tests/software/gatk4/bedtointervallist -entry test_gatk4_bedtointervallist -c tests/config/nextflow.config
+- name: gatk4 bedtointervallist test_gatk4_bedtointervallist
+  command: nextflow run tests/software/gatk4/bedtointervallist -entry test_gatk4_bedtointervallist -c tests/config/nextflow.config
   tags:
     - gatk4
     - gatk4_bedtointervallist
diff --git a/tests/software/gatk4/createsequencedictionary/main.nf b/tests/software/gatk4/createsequencedictionary/main.nf
index 5dfeb8cb..ff798693 100644
--- a/tests/software/gatk4/createsequencedictionary/main.nf
+++ b/tests/software/gatk4/createsequencedictionary/main.nf
@@ -5,7 +5,7 @@ nextflow.enable.dsl = 2
 include { GATK4_CREATESEQUENCEDICTIONARY } from '../../../../software/gatk4/createsequencedictionary/main.nf' addParams( options: [:] )
 
 workflow test_gatk4_createsequencedictionary {
-    fasta = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
+    fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
 
     GATK4_CREATESEQUENCEDICTIONARY ( fasta )
 }
diff --git a/tests/software/gatk4/createsequencedictionary/test.yml b/tests/software/gatk4/createsequencedictionary/test.yml
index bbee03d7..4b849d5a 100644
--- a/tests/software/gatk4/createsequencedictionary/test.yml
+++ b/tests/software/gatk4/createsequencedictionary/test.yml
@@ -1,8 +1,8 @@
-- name: gatk4 createsequencedictionary
-  command: nextflow run ./tests/software/gatk4/createsequencedictionary -entry test_gatk4_createsequencedictionary -c tests/config/nextflow.config
+- name: gatk4 createsequencedictionary test_gatk4_createsequencedictionary
+  command: nextflow run tests/software/gatk4/createsequencedictionary -entry test_gatk4_createsequencedictionary -c tests/config/nextflow.config
   tags:
     - gatk4
     - gatk4_createsequencedictionary
   files:
-    - path: ./output/gatk4/test_genome.dict
-      md5sum: 829dd07bd0c9d3be2fd060238c0b37c3
+    - path: output/gatk4/genome.dict
+      md5sum: 7362679f176e0f52add03c08f457f646
diff --git a/tests/software/gatk4/fastqtosam/main.nf b/tests/software/gatk4/fastqtosam/main.nf
index a9d11295..82ddcb99 100644
--- a/tests/software/gatk4/fastqtosam/main.nf
+++ b/tests/software/gatk4/fastqtosam/main.nf
@@ -6,7 +6,7 @@ include { GATK4_FASTQTOSAM } from '../../../../software/gatk4/fastqtosam/main.nf
 
 workflow test_gatk4_fastqtosam_single_end {
     input = [ [ id:'test', single_end:true ], // meta map
-              file("${launchDir}/tests/data/genomics/sarscov2/illumina/fastq/test_1.fastq.gz", checkIfExists: true) 
+              file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
             ]
 
     GATK4_FASTQTOSAM ( input )
@@ -14,8 +14,8 @@ workflow test_gatk4_fastqtosam_single_end {
 
 workflow test_gatk4_fastqtosam_paired_end {
     input = [ [ id:'test', single_end:false ], // meta map
-              file("${launchDir}/tests/data/genomics/sarscov2/illumina/fastq/test_1.fastq.gz", checkIfExists: true),
-              file("${launchDir}/tests/data/genomics/sarscov2/illumina/fastq/test_2.fastq.gz", checkIfExists: true) 
+              file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true),
+              file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true)
             ]
 
     GATK4_FASTQTOSAM ( input )
diff --git a/tests/software/gatk4/fastqtosam/test.yml b/tests/software/gatk4/fastqtosam/test.yml
index 7d4926b4..2674ab59 100644
--- a/tests/software/gatk4/fastqtosam/test.yml
+++ b/tests/software/gatk4/fastqtosam/test.yml
@@ -2,8 +2,8 @@
   command: nextflow run tests/software/gatk4/fastqtosam -entry test_gatk4_fastqtosam_single_end -c tests/config/nextflow.config
   tags:
     - gatk4_fastqtosam_single_end
-    - gatk4_fastqtosam
     - gatk4
+    - gatk4_fastqtosam
   files:
     - path: output/gatk4/test.bam
       md5sum: 4967100b2e4912c0e4ce0976d946bafb
@@ -12,8 +12,8 @@
   command: nextflow run tests/software/gatk4/fastqtosam -entry test_gatk4_fastqtosam_paired_end -c tests/config/nextflow.config
   tags:
     - gatk4_fastqtosam_paired_end
-    - gatk4_fastqtosam
     - gatk4
+    - gatk4_fastqtosam
   files:
     - path: output/gatk4/test.bam
       md5sum: 4967100b2e4912c0e4ce0976d946bafb
diff --git a/tests/software/gatk4/haplotypecaller/main.nf b/tests/software/gatk4/haplotypecaller/main.nf
new file mode 100644
index 00000000..9e77d6aa
--- /dev/null
+++ b/tests/software/gatk4/haplotypecaller/main.nf
@@ -0,0 +1,17 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_HAPLOTYPECALLER } from '../../../../software/gatk4/haplotypecaller/main.nf' addParams( options: [:] )
+
+workflow test_gatk4_haplotypecaller {
+    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)
+
+    GATK4_HAPLOTYPECALLER ( input, fasta, fai, dict )
+}
diff --git a/tests/software/gatk4/haplotypecaller/test.yml b/tests/software/gatk4/haplotypecaller/test.yml
new file mode 100644
index 00000000..42fd9e18
--- /dev/null
+++ b/tests/software/gatk4/haplotypecaller/test.yml
@@ -0,0 +1,13 @@
+- name: gatk4 haplotypecaller test_gatk4_haplotypecaller
+  command: nextflow run tests/software/gatk4/haplotypecaller -entry test_gatk4_haplotypecaller -c tests/config/nextflow.config
+  tags:
+    - gatk4_haplotypecaller
+    - gatk4
+  files:
+    - path: output/gatk4/test.vcf.gz
+      should_exist: true
+      contains:
+        - 'MT192765.1'
+        - '54.60'
+        - '37.32'
+    - path: output/gatk4/test.vcf.gz.tbi
diff --git a/tests/software/gatk4/mergebamalignment/main.nf b/tests/software/gatk4/mergebamalignment/main.nf
index 31c8a99e..302f9db3 100644
--- a/tests/software/gatk4/mergebamalignment/main.nf
+++ b/tests/software/gatk4/mergebamalignment/main.nf
@@ -6,11 +6,11 @@ include { GATK4_MERGEBAMALIGNMENT } from '../../../../software/gatk4/mergebamali
 
 workflow test_gatk4_mergebamalignment {
     input    = [ [ id:'test' ], // meta map
-                 file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_single_end.bam", checkIfExists: true) 
+                 file(params.test_data['sarscov2']['illumina']['test_single_end_bam'], checkIfExists: true)
                ]
-    unmapped = file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_unaligned.bam", checkIfExists: true)
-    fasta    = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true)
-    dict     = file("${launchDir}/tests/data/genomics/sarscov2/genome/genome.dict", checkIfExists: true)
+    unmapped = file(params.test_data['sarscov2']['illumina']['test_unaligned_bam'], checkIfExists: true)
+    fasta    = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    dict     = file(params.test_data['sarscov2']['genome']['genome_dict'], checkIfExists: true)
 
     GATK4_MERGEBAMALIGNMENT ( input, unmapped, fasta, dict )
 }
diff --git a/tests/software/gatk4/mergebamalignment/test.yml b/tests/software/gatk4/mergebamalignment/test.yml
index 088cd5d9..2fb1be5f 100644
--- a/tests/software/gatk4/mergebamalignment/test.yml
+++ b/tests/software/gatk4/mergebamalignment/test.yml
@@ -1,8 +1,8 @@
-- name: gatk4 mergebamalignment
-  command: nextflow run ./tests/software/gatk4/mergebamalignment -entry test_gatk4_mergebamalignment -c tests/config/nextflow.config
+- name: gatk4 mergebamalignment test_gatk4_mergebamalignment
+  command: nextflow run tests/software/gatk4/mergebamalignment -entry test_gatk4_mergebamalignment -c tests/config/nextflow.config
   tags:
     - gatk4
     - gatk4_mergebamalignment
   files:
     - path: output/gatk4/test.merged.bam
-      md5sum: 825816a6b288d644318c46cc9968e129
+      md5sum: bd4a5e2ea916826aadebb5878333e26f
diff --git a/tests/software/gatk4/mergevcfs/main.nf b/tests/software/gatk4/mergevcfs/main.nf
index c7a3345e..26549c34 100644
--- a/tests/software/gatk4/mergevcfs/main.nf
+++ b/tests/software/gatk4/mergevcfs/main.nf
@@ -6,10 +6,10 @@ include { GATK4_MERGEVCFS } from '../../../../software/gatk4/mergevcfs/main.nf'
 
 workflow test_gatk4_mergevcfs {
     input = [ [ id:'test' ], // meta map
-              [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf", checkIfExists: true),
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test2.vcf", checkIfExists: true) ] 
+              [ file(params.test_data['sarscov2']['illumina']['test_vcf'], checkIfExists: true),
+                file(params.test_data['sarscov2']['illumina']['test2_vcf'], checkIfExists: true) ] 
             ]
-    dict = file('tests/data/genomics/sarscov2/genome/genome.dict', checkIfExists: true)
+    dict = file(params.test_data['sarscov2']['genome']['genome_dict'], checkIfExists: true)
 
     GATK4_MERGEVCFS ( input, dict, false )
 }
@@ -17,10 +17,10 @@ workflow test_gatk4_mergevcfs {
 workflow test_gatk4_mergevcfs_refdict {
     def input = []
     input = [ [ id:'test' ], // meta map
-              [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test.vcf", checkIfExists: true),
-                file("${launchDir}/tests/data/genomics/sarscov2/illumina/vcf/test2.vcf", checkIfExists: true) ] 
+              [ file(params.test_data['sarscov2']['illumina']['test_vcf'], checkIfExists: true),
+                file(params.test_data['sarscov2']['illumina']['test2_vcf'], checkIfExists: true) ] 
             ]
-    dict  = file('tests/data/genomics/sarscov2/genome/genome.dict', checkIfExists: true)
+    dict  = file(params.test_data['sarscov2']['genome']['genome_dict'], checkIfExists: true)
 
-    GATK4_MERGEVCFS ( input, ref_dict, true )
+    GATK4_MERGEVCFS ( input, dict, true )
 }
diff --git a/tests/software/gatk4/mergevcfs/test.yml b/tests/software/gatk4/mergevcfs/test.yml
index 2bd2e50c..c461710b 100644
--- a/tests/software/gatk4/mergevcfs/test.yml
+++ b/tests/software/gatk4/mergevcfs/test.yml
@@ -1,17 +1,18 @@
-- name: gatk4 mergevcfs
-  command: nextflow run ./tests/software/gatk4/mergevcfs -entry test_gatk4_mergevcfs -c tests/config/nextflow.config
+- name: gatk4 mergevcfs test_gatk4_mergevcfs
+  command: nextflow run tests/software/gatk4/mergevcfs -entry test_gatk4_mergevcfs -c tests/config/nextflow.config
   tags:
-    - gatk4
     - gatk4_mergevcfs
+    - gatk4
   files:
     - path: output/gatk4/test.merged.vcf.gz
       md5sum: ff48f175e26db2d4b2957762f6d1c715
 
-- name: gatk4 mergevcfs refdict
-  command: nextflow run ./tests/software/gatk4/mergevcfs -entry test_gatk4_mergevcfs_refdict -c tests/config/nextflow.config
+- name: gatk4 mergevcfs test_gatk4_mergevcfs_refdict
+  command: nextflow run tests/software/gatk4/mergevcfs -entry test_gatk4_mergevcfs_refdict -c tests/config/nextflow.config
   tags:
-    - gatk4
     - gatk4_mergevcfs
+    - gatk4
+    - gatk4_mergevcfs_refdict
   files:
     - path: output/gatk4/test.merged.vcf.gz
       md5sum: ff48f175e26db2d4b2957762f6d1c715
diff --git a/tests/software/gatk4/revertsam/main.nf b/tests/software/gatk4/revertsam/main.nf
index b452d0b5..2bc45f74 100644
--- a/tests/software/gatk4/revertsam/main.nf
+++ b/tests/software/gatk4/revertsam/main.nf
@@ -6,7 +6,7 @@ include { GATK4_REVERTSAM } from '../../../../software/gatk4/revertsam/main.nf'
 
 workflow test_gatk4_revertsam {
     input = [ [ id:'test' ], // meta map
-              file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.bam", checkIfExists: true) 
+              file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true)
             ]
 
     GATK4_REVERTSAM ( input )
diff --git a/tests/software/gatk4/revertsam/test.yml b/tests/software/gatk4/revertsam/test.yml
index 7689af98..221975bf 100644
--- a/tests/software/gatk4/revertsam/test.yml
+++ b/tests/software/gatk4/revertsam/test.yml
@@ -1,8 +1,8 @@
-- name: gatk4 revertsam
-  command: nextflow run ./tests/software/gatk4/revertsam -entry test_gatk4_revertsam -c tests/config/nextflow.config
+- name: gatk4 revertsam test_gatk4_revertsam
+  command: nextflow run tests/software/gatk4/revertsam -entry test_gatk4_revertsam -c tests/config/nextflow.config
   tags:
     - gatk4
     - gatk4_revertsam
   files:
     - path: output/gatk4/test.reverted.bam
-      md5sum: 6de199f7a0577e2a3fe7b3a9cb4ae25d
+      md5sum: f778310b18b83b49929eb648594f96dc
diff --git a/tests/software/gatk4/samtofastq/main.nf b/tests/software/gatk4/samtofastq/main.nf
index 1fb58138..9c3e5341 100644
--- a/tests/software/gatk4/samtofastq/main.nf
+++ b/tests/software/gatk4/samtofastq/main.nf
@@ -6,7 +6,7 @@ include { GATK4_SAMTOFASTQ } from '../../../../software/gatk4/samtofastq/main.nf
 
 workflow test_gatk4_samtofastq_single_end {
     input = [ [ id:'test', single_end: true ], // meta map
-              [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_single_end.bam", checkIfExists: true) ] 
+              [ file(params.test_data['sarscov2']['illumina']['test_single_end_bam'], checkIfExists: true) ]
             ]
 
     GATK4_SAMTOFASTQ ( input )
@@ -14,7 +14,7 @@ workflow test_gatk4_samtofastq_single_end {
 
 workflow test_gatk4_samtofastq_paired_end {
     input = [ [ id:'test', single_end: false ], // meta map
-              [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.bam", checkIfExists: true) ] 
+              [ file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true) ]
             ]
 
     GATK4_SAMTOFASTQ ( input )
diff --git a/tests/software/gatk4/samtofastq/test.yml b/tests/software/gatk4/samtofastq/test.yml
index e47ac6eb..a2a900a1 100644
--- a/tests/software/gatk4/samtofastq/test.yml
+++ b/tests/software/gatk4/samtofastq/test.yml
@@ -1,21 +1,21 @@
-- name: gatk4 samtofastq single-end
-  command: nextflow run ./tests/software/gatk4/samtofastq -entry test_gatk4_samtofastq_single_end -c tests/config/nextflow.config
+- name: gatk4 samtofastq test_gatk4_samtofastq_single_end
+  command: nextflow run tests/software/gatk4/samtofastq -entry test_gatk4_samtofastq_single_end -c tests/config/nextflow.config
   tags:
     - gatk4
     - gatk4_samtofastq
     - gatk4_samtofastq_single_end
   files:
-    - path: ./output/gatk4/test.fastq.gz
+    - path: output/gatk4/test.fastq.gz
       md5sum: 50ace41d4c24467f24f8b929540a7797
 
-- name: gatk4 samtofastq paired-end
-  command: nextflow run ./tests/software/gatk4/samtofastq -entry test_gatk4_samtofastq_paired_end -c tests/config/nextflow.config
+- name: gatk4 samtofastq test_gatk4_samtofastq_paired_end
+  command: nextflow run tests/software/gatk4/samtofastq -entry test_gatk4_samtofastq_paired_end -c tests/config/nextflow.config
   tags:
     - gatk4
     - gatk4_samtofastq
     - gatk4_samtofastq_paired_end
   files:
-    - path: ./output/gatk4/test_2.fastq.gz
-      md5sum: 613bf64c023609e1c62ad6ce9e4be8d7
-    - path: ./output/gatk4/test_1.fastq.gz
+    - path: output/gatk4/test_1.fastq.gz
       md5sum: cfea607c9d75fd9ea9704780ad3a499c
+    - path: output/gatk4/test_2.fastq.gz
+      md5sum: 613bf64c023609e1c62ad6ce9e4be8d7
diff --git a/tests/software/gatk4/splitncigarreads/main.nf b/tests/software/gatk4/splitncigarreads/main.nf
index 5c3848f7..8dba64cb 100644
--- a/tests/software/gatk4/splitncigarreads/main.nf
+++ b/tests/software/gatk4/splitncigarreads/main.nf
@@ -6,11 +6,11 @@ include { GATK4_SPLITNCIGARREADS } from '../../../../software/gatk4/splitncigarr
 
 workflow test_gatk4_splitncigarreads {
     input = [ [ id:'test' ], // meta map
-              [ file("${launchDir}/tests/data/genomics/sarscov2/illumina/bam/test_paired_end.bam", checkIfExists: true)] 
+              [ file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true) ] 
             ]
-    fasta = [ file("tests/data/genomics/sarscov2/genome/genome.fasta", checkIfExists: true),
-              file("tests/data/genomics/sarscov2/genome/genome.fasta.fai", checkIfExists: true),
-              file("tests/data/genomics/sarscov2/genome/genome.dict", checkIfExists: true)
+    fasta = [ 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)
             ]
 
     GATK4_SPLITNCIGARREADS ( input, fasta )
diff --git a/tests/software/gatk4/splitncigarreads/test.yml b/tests/software/gatk4/splitncigarreads/test.yml
index 4afa9fcf..8f08899d 100644
--- a/tests/software/gatk4/splitncigarreads/test.yml
+++ b/tests/software/gatk4/splitncigarreads/test.yml
@@ -1,8 +1,8 @@
-- name: gatk4 splitncigarreads
-  command: nextflow run ./tests/software/gatk4/splitncigarreads -entry test_gatk4_splitncigarreads -c tests/config/nextflow.config
+- name: gatk4 splitncigarreads test_gatk4_splitncigarreads
+  command: nextflow run tests/software/gatk4/splitncigarreads -entry test_gatk4_splitncigarreads -c tests/config/nextflow.config
   tags:
     - gatk4
     - gatk4_splitncigarreads
   files:
     - path: output/gatk4/test.split_cigar.bam
-      md5sum: 9dd44d3988b9ff238a6b4771cdf13eba
+      md5sum: a8473bd4abc7b13751fc51b92531da4c