From f65d896f57a5564a40278f4df7f627e8f8a17d6b Mon Sep 17 00:00:00 2001
From: Mahesh Binzer-Panchal <mahesh.binzer-panchal@nbis.se>
Date: Mon, 9 May 2022 15:09:15 +0000
Subject: [PATCH 01/16] create genomescope2 module

---
 modules/genomescope2/main.nf               | 42 +++++++++++++++++++++
 modules/genomescope2/meta.yml              | 43 ++++++++++++++++++++++
 tests/config/pytest_modules.yml            | 12 ++++--
 tests/modules/genomescope2/main.nf         | 19 ++++++++++
 tests/modules/genomescope2/nextflow.config | 13 +++++++
 tests/modules/genomescope2/test.yml        | 12 ++++++
 6 files changed, 137 insertions(+), 4 deletions(-)
 create mode 100644 modules/genomescope2/main.nf
 create mode 100644 modules/genomescope2/meta.yml
 create mode 100644 tests/modules/genomescope2/main.nf
 create mode 100644 tests/modules/genomescope2/nextflow.config
 create mode 100644 tests/modules/genomescope2/test.yml

diff --git a/modules/genomescope2/main.nf b/modules/genomescope2/main.nf
new file mode 100644
index 00000000..d81ed27a
--- /dev/null
+++ b/modules/genomescope2/main.nf
@@ -0,0 +1,42 @@
+process GENOMESCOPE2 {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::genomescope2=2.0" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/genomescope2:2.0--py310r41hdfd78af_5':
+        'quay.io/biocontainers/genomescope2:2.0--py310r41hdfd78af_5' }"
+
+    input:
+    tuple val(meta), path(histogram)
+
+    output:
+    tuple val(meta), path("$prefix-linear_plot.png")            , emit: linear_plot_png
+    tuple val(meta), path("$prefix-transformed_linear_plot.png"), emit: transformed_linear_plot_png
+    tuple val(meta), path("$prefix-log_plot.png")               , emit: log_plot_png
+    tuple val(meta), path("$prefix-transformed_log_plot.png")   , emit: transformed_log_plot_png
+    tuple val(meta), path("$prefix-model.txt")                  , emit: model
+    tuple val(meta), path("$prefix-summary.txt")                , emit: summary
+    path "versions.yml"                                         , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    genomescope2 \\
+        --input $histogram \\
+        $args \\
+        --output . \\
+        --name_prefix $prefix
+
+    ls -l
+
+    cat <<-END_VERSIONS > versions.yml
+    '${task.process}':
+        genomescope2: \$( genomescope2 -v | sed 's/GenomeScope //' )
+    END_VERSIONS
+    """
+}
diff --git a/modules/genomescope2/meta.yml b/modules/genomescope2/meta.yml
new file mode 100644
index 00000000..b680bb8f
--- /dev/null
+++ b/modules/genomescope2/meta.yml
@@ -0,0 +1,43 @@
+name: "genomescope2"
+description: Estimate genome heterozygosity, repeat content, and size from sequencing reads using a kmer-based statistical approach
+keywords:
+  - "genome size"
+  - "genome heterozygosity"
+  - "repeat content"
+tools:
+  - "genomescope2":
+      description: "Reference-free profiling of polyploid genomes"
+      homepage: "http://qb.cshl.edu/genomescope/genomescope2.0/"
+      documentation: "https://github.com/tbenavi1/genomescope2.0/blob/master/README.md"
+      tool_dev_url: "https://github.com/tbenavi1/genomescope2.0"
+      doi: "https://doi.org/10.1038/s41467-020-14998-3"
+      licence: "['Apache License, Version 2.0 (Apache-2.0)']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - histogram:
+      type: file
+      description: A K-mer histogram file
+      pattern: "*.hist"
+
+output:
+  - 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"
+  - linear_plot_png:
+      type: file
+      description: A genomescope2 linear plot in PNG format
+      pattern: "*.png"
+
+authors:
+  - "@mahesh-panchal"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index ba292bce..286bd8b8 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -835,6 +835,10 @@ genmap/mappability:
   - modules/genmap/mappability/**
   - tests/modules/genmap/mappability/**
 
+genomescope2:
+  - modules/genomescope2/**
+  - tests/modules/genomescope2/**
+
 genrich:
   - modules/genrich/**
   - tests/modules/genrich/**
@@ -1639,14 +1643,14 @@ samtools/bam2fq:
   - modules/samtools/bam2fq/**
   - tests/modules/samtools/bam2fq/**
 
-samtools/convert:
-  - modules/samtools/convert/**
-  - tests/modules/samtools/convert/**
-
 samtools/collatefastq:
   - modules/samtools/collatefastq/**
   - tests/modules/samtools/collatefastq/**
 
+samtools/convert:
+  - modules/samtools/convert/**
+  - tests/modules/samtools/convert/**
+
 samtools/depth:
   - modules/samtools/depth/**
   - tests/modules/samtools/depth/**
diff --git a/tests/modules/genomescope2/main.nf b/tests/modules/genomescope2/main.nf
new file mode 100644
index 00000000..9f15be88
--- /dev/null
+++ b/tests/modules/genomescope2/main.nf
@@ -0,0 +1,19 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { MERYL_COUNT     } from '../../../modules/meryl/count/main.nf'
+include { MERYL_HISTOGRAM } from '../../../modules/meryl/histogram/main.nf'
+include { GENOMESCOPE2    } from '../../../modules/genomescope2/main.nf'
+
+workflow test_genomescope2 {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+    ]
+
+    MERYL_COUNT ( input )
+    MERYL_HISTOGRAM ( MERYL_COUNT.out.meryl_db )
+    GENOMESCOPE2 ( MERYL_HISTOGRAM.out.hist )
+}
diff --git a/tests/modules/genomescope2/nextflow.config b/tests/modules/genomescope2/nextflow.config
new file mode 100644
index 00000000..29a0be3a
--- /dev/null
+++ b/tests/modules/genomescope2/nextflow.config
@@ -0,0 +1,13 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: 'MERYL.*' {
+        ext.args = 'k=21'
+    }
+
+    withName: 'GENOMESCOPE2' {
+        ext.args = '-k 21 -p 1'
+    }
+
+}
diff --git a/tests/modules/genomescope2/test.yml b/tests/modules/genomescope2/test.yml
new file mode 100644
index 00000000..d2a664d9
--- /dev/null
+++ b/tests/modules/genomescope2/test.yml
@@ -0,0 +1,12 @@
+## TODO nf-core: Please run the following command to build this file:
+#                nf-core modules create-test-yml genomescope2
+- name: "genomescope2"
+  command: nextflow run ./tests/modules/genomescope2 -entry test_genomescope2 -c ./tests/config/nextflow.config -c ./tests/modules/genomescope2/nextflow.config
+  tags:
+    - "genomescope2"
+    #
+  files:
+    - path: "output/genomescope2/test.bam"
+      md5sum: e667c7caad0bc4b7ac383fd023c654fc
+    - path: output/genomescope2/versions.yml
+      md5sum: a01fe51bc4c6a3a6226fbf77b2c7cf3b

From d54cf467f47461e8304e6fc4a4a1192869bfd792 Mon Sep 17 00:00:00 2001
From: Mahesh Binzer-Panchal <mahesh.binzer-panchal@nbis.se>
Date: Wed, 11 May 2022 10:42:17 +0000
Subject: [PATCH 02/16] Update outputs and test data set

---
 modules/genomescope2/main.nf        | 16 +++++++---------
 modules/genomescope2/meta.yml       | 26 +++++++++++++++++++++++++-
 tests/modules/genomescope2/main.nf  |  2 +-
 tests/modules/genomescope2/test.yml | 29 ++++++++++++++++++++---------
 4 files changed, 53 insertions(+), 20 deletions(-)

diff --git a/modules/genomescope2/main.nf b/modules/genomescope2/main.nf
index d81ed27a..2ddf9e43 100644
--- a/modules/genomescope2/main.nf
+++ b/modules/genomescope2/main.nf
@@ -11,13 +11,13 @@ process GENOMESCOPE2 {
     tuple val(meta), path(histogram)
 
     output:
-    tuple val(meta), path("$prefix-linear_plot.png")            , emit: linear_plot_png
-    tuple val(meta), path("$prefix-transformed_linear_plot.png"), emit: transformed_linear_plot_png
-    tuple val(meta), path("$prefix-log_plot.png")               , emit: log_plot_png
-    tuple val(meta), path("$prefix-transformed_log_plot.png")   , emit: transformed_log_plot_png
-    tuple val(meta), path("$prefix-model.txt")                  , emit: model
-    tuple val(meta), path("$prefix-summary.txt")                , emit: summary
-    path "versions.yml"                                         , emit: versions
+    tuple val(meta), path("*_linear_plot.png")            , emit: linear_plot_png
+    tuple val(meta), path("*_transformed_linear_plot.png"), emit: transformed_linear_plot_png
+    tuple val(meta), path("*_log_plot.png")               , emit: log_plot_png
+    tuple val(meta), path("*_transformed_log_plot.png")   , emit: transformed_log_plot_png
+    tuple val(meta), path("*_model.txt")                  , emit: model
+    tuple val(meta), path("*_summary.txt")                , emit: summary
+    path "versions.yml"                                   , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
@@ -32,8 +32,6 @@ process GENOMESCOPE2 {
         --output . \\
         --name_prefix $prefix
 
-    ls -l
-
     cat <<-END_VERSIONS > versions.yml
     '${task.process}':
         genomescope2: \$( genomescope2 -v | sed 's/GenomeScope //' )
diff --git a/modules/genomescope2/meta.yml b/modules/genomescope2/meta.yml
index b680bb8f..505daafe 100644
--- a/modules/genomescope2/meta.yml
+++ b/modules/genomescope2/meta.yml
@@ -37,7 +37,31 @@ output:
   - linear_plot_png:
       type: file
       description: A genomescope2 linear plot in PNG format
-      pattern: "*.png"
+      pattern: "*_linear_plot.png"
+  - linear_plot_png:
+      type: file
+      description: A genomescope2 linear plot in PNG format
+      pattern: "*_linear_plot.png"
+  - transformed_linear_plot_png:
+      type: file
+      description: A genomescope2 transformed linear plot in PNG format
+      pattern: "*_transformed_linear_plot.png"
+  - log_plot_png:
+      type: file
+      description: A genomescope2 log plot in PNG format
+      pattern: "*_log_plot.png"
+  - transformed_log_plot_png:
+      type: file
+      description: A genomescope2 transformed log plot in PNG format
+      pattern: "*_transformed_log_plot.png"
+  - model:
+      type: file
+      description: Genomescope2 model fit summary
+      pattern: "*_model.txt"
+  - summary:
+      type: file
+      description: Genomescope2 histogram summary
+      pattern: "*_summary.txt"
 
 authors:
   - "@mahesh-panchal"
diff --git a/tests/modules/genomescope2/main.nf b/tests/modules/genomescope2/main.nf
index 9f15be88..5ceebfd6 100644
--- a/tests/modules/genomescope2/main.nf
+++ b/tests/modules/genomescope2/main.nf
@@ -10,7 +10,7 @@ workflow test_genomescope2 {
 
     input = [
         [ id:'test', single_end:false ], // meta map
-        file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+        file(params.test_data['bacteroides_fragilis']['illumina']['test1_1_fastq_gz'], checkIfExists: true)
     ]
 
     MERYL_COUNT ( input )
diff --git a/tests/modules/genomescope2/test.yml b/tests/modules/genomescope2/test.yml
index d2a664d9..accd2a14 100644
--- a/tests/modules/genomescope2/test.yml
+++ b/tests/modules/genomescope2/test.yml
@@ -1,12 +1,23 @@
-## TODO nf-core: Please run the following command to build this file:
-#                nf-core modules create-test-yml genomescope2
-- name: "genomescope2"
-  command: nextflow run ./tests/modules/genomescope2 -entry test_genomescope2 -c ./tests/config/nextflow.config -c ./tests/modules/genomescope2/nextflow.config
+- name: genomescope2 test_genomescope2
+  command: nextflow run tests/modules/genomescope2 -entry test_genomescope2 -c tests/config/nextflow.config
   tags:
-    - "genomescope2"
-    #
+    - genomescope2
   files:
-    - path: "output/genomescope2/test.bam"
-      md5sum: e667c7caad0bc4b7ac383fd023c654fc
+    - path: output/genomescope2/test_linear_plot.png
+      md5sum: 94c165c5028156299a1d4d05766cac51
+    - path: output/genomescope2/test_log_plot.png
+      md5sum: 9d25ca463d92a0c73a893da7fd3979ba
+    - path: output/genomescope2/test_model.txt
+      md5sum: 3caf62f715f64a2f2b8fdff5d079cb84
+    - path: output/genomescope2/test_summary.txt
+      md5sum: 7452860e2cea99b85f3ff60daeac77f5
+    - path: output/genomescope2/test_transformed_linear_plot.png
+      md5sum: 99a64c1c18d8670f64cb863d4334abbb
+    - path: output/genomescope2/test_transformed_log_plot.png
+      md5sum: b4e029c9fb9987ca33b17392a691c1b4
     - path: output/genomescope2/versions.yml
-      md5sum: a01fe51bc4c6a3a6226fbf77b2c7cf3b
+      md5sum: 18afeb26f62a47f680b2bb3e27da9cbc
+    - path: output/meryl/test.hist
+      md5sum: f75362ab9cd70d96621b3690e952085f
+    - path: output/meryl/versions.yml
+      md5sum: 944b7ea81a82bd20174c5042857003fc

From 90c67794bfe33cd82954ca482979d5e8d5d3fc12 Mon Sep 17 00:00:00 2001
From: Chris Cheshire <chris.j.cheshire@gmail.com>
Date: Thu, 12 May 2022 10:01:33 +0100
Subject: [PATCH 03/16] Added extra tests but error with collection

---
 modules/bowtie2/align/main.nf               |  1 -
 tests/modules/bowtie2/align/main.nf         | 29 +++++++++++
 tests/modules/bowtie2/align/nextflow.config |  1 +
 tests/modules/bowtie2/align/test.yml        | 56 +++++++++++++++------
 4 files changed, 71 insertions(+), 16 deletions(-)

diff --git a/modules/bowtie2/align/main.nf b/modules/bowtie2/align/main.nf
index 38242fcb..e4bb4327 100644
--- a/modules/bowtie2/align/main.nf
+++ b/modules/bowtie2/align/main.nf
@@ -69,4 +69,3 @@ process BOWTIE2_ALIGN {
     END_VERSIONS
     """
 }
-
diff --git a/tests/modules/bowtie2/align/main.nf b/tests/modules/bowtie2/align/main.nf
index f91394ff..42e2306a 100644
--- a/tests/modules/bowtie2/align/main.nf
+++ b/tests/modules/bowtie2/align/main.nf
@@ -30,6 +30,35 @@ workflow test_bowtie2_align_paired_end {
     fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
     save_unaligned = false
 
+    BOWTIE2_BUILD ( fasta )
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned )
+}
+
+workflow test_bowtie2_align_single_end_large_index {
+    input = [
+        [ id:'test', single_end:true ], // meta map
+        [
+            file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+        ]
+    ]
+    fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    save_unaligned = false
+
+    BOWTIE2_BUILD ( fasta )
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned )
+}
+
+workflow test_bowtie2_align_paired_end_large_index {
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        [
+            file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true),
+            file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true)
+        ]
+    ]
+    fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    save_unaligned = false
+
     BOWTIE2_BUILD ( fasta )
     BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned )
 }
\ No newline at end of file
diff --git a/tests/modules/bowtie2/align/nextflow.config b/tests/modules/bowtie2/align/nextflow.config
index b4640de7..08f7bed0 100644
--- a/tests/modules/bowtie2/align/nextflow.config
+++ b/tests/modules/bowtie2/align/nextflow.config
@@ -5,6 +5,7 @@ params {
 process {
 
     publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
 }
 
 if (params.force_large_index) {
diff --git a/tests/modules/bowtie2/align/test.yml b/tests/modules/bowtie2/align/test.yml
index ef05d70d..8a33c928 100644
--- a/tests/modules/bowtie2/align/test.yml
+++ b/tests/modules/bowtie2/align/test.yml
@@ -1,21 +1,47 @@
 - name: bowtie2 align test_bowtie2_align_single_end
   command: nextflow run ./tests/modules/bowtie2/align -entry test_bowtie2_align_single_end -c ./tests/config/nextflow.config -c ./tests/modules/bowtie2/align/nextflow.config
   tags:
-    - bowtie2/align
     - bowtie2
+    - bowtie2/align
   files:
-    - path: output/bowtie2/test.bam
-    - path: output/bowtie2/test.bowtie2.log
-      md5sum: 7b8a9e61b7646da1089b041333c41a87
-    - path: output/bowtie2/versions.yml
+    - path: ./output/bowtie2/test.bam
+    # - path: output/bowtie2/test.bowtie2.log
+    #   md5sum: 7b8a9e61b7646da1089b041333c41a87
+    # - path: output/bowtie2/versions.yml
+    #   md5sum: 24621c58884fe90c2255ccd1fe4352ae
 
-- name: bowtie2 align test_bowtie2_align_paired_end
-  command: nextflow run ./tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end -c ./tests/config/nextflow.config -c ./tests/modules/bowtie2/align/nextflow.config
-  tags:
-    - bowtie2/align
-    - bowtie2
-  files:
-    - path: output/bowtie2/test.bam
-    - path: output/bowtie2/test.bowtie2.log
-      md5sum: bd89ce1b28c93bf822bae391ffcedd19
-    - path: output/bowtie2/versions.yml
+# - name: bowtie2 align test_bowtie2_align_paired_end
+#   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config
+#   tags:
+#     - bowtie2
+#     - bowtie2/align
+#   files:
+#     - path: output/bowtie2/test.bam
+#     - path: output/bowtie2/test.bowtie2.log
+#       md5sum: bd89ce1b28c93bf822bae391ffcedd19
+#     - path: output/bowtie2/versions.yml
+#       md5sum: f9712ca6d75393ad5c1781a2dcf82d32
+
+# - name: bowtie2 align test_bowtie2_align_single_end_large_index
+#   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_single_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
+#   tags:
+#     - bowtie2
+#     - bowtie2/align
+#   files:
+#     - path: output/bowtie2/test.bam
+#     - path: output/bowtie2/test.bowtie2.log
+#       md5sum: 7b8a9e61b7646da1089b041333c41a87
+#     - path: output/bowtie2/versions.yml
+#       md5sum: bf0537964a85ce9461ae1b0e2f260211
+
+# - name: bowtie2 align test_bowtie2_align_paired_end_large_index
+#   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
+#   tags:
+#     - bowtie2
+#     - bowtie2/align
+#   files:
+#     - path: output/bowtie2/test.bam
+#     - path: output/bowtie2/test.bowtie2.log
+#       md5sum: bd89ce1b28c93bf822bae391ffcedd19
+#     - path: output/bowtie2/versions.yml
+#       md5sum: c0a88953500eaf46de7ff378a7de4a37

From 6fa1c8089afe83e612dbc14e4b85f277166c4a2b Mon Sep 17 00:00:00 2001
From: Nicolas Vannieuwkerke <nicolas.vannieuwkerke@ugent.be>
Date: Fri, 13 May 2022 10:03:20 +0200
Subject: [PATCH 04/16] Added memory allocation to VCFeval

---
 modules/rtgtools/vcfeval/main.nf | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/modules/rtgtools/vcfeval/main.nf b/modules/rtgtools/vcfeval/main.nf
index 1bad4231..27a488f7 100644
--- a/modules/rtgtools/vcfeval/main.nf
+++ b/modules/rtgtools/vcfeval/main.nf
@@ -35,12 +35,13 @@ process RTGTOOLS_VCFEVAL {
     def eval_regions = evaluation_regions ? "--evaluation-regions=$evaluation_regions" : ""
     def truth_index = truth_vcf_tbi ? "" : "rtg index $truth_vcf"
     def query_index = query_vcf_tbi ? "" : "rtg index $query_vcf"
+    def avail_mem = task.memory.toGiga() + "G"
 
     """
     $truth_index
     $query_index
 
-    rtg vcfeval \\
+    rtg RTG_MEM=$avail_mem vcfeval \\
         $args \\
         --baseline=$truth_vcf \\
         $bed_regions \\

From 5df0bcf1bed5e5fd4e2d1850694e386e0b85b240 Mon Sep 17 00:00:00 2001
From: Mahesh Binzer-Panchal <mahesh.binzer-panchal@nbis.se>
Date: Fri, 13 May 2022 11:27:42 +0200
Subject: [PATCH 05/16] Update tests/modules/genomescope2/test.yml

Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de>
---
 tests/modules/genomescope2/test.yml | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tests/modules/genomescope2/test.yml b/tests/modules/genomescope2/test.yml
index accd2a14..5b03f545 100644
--- a/tests/modules/genomescope2/test.yml
+++ b/tests/modules/genomescope2/test.yml
@@ -20,4 +20,3 @@
     - path: output/meryl/test.hist
       md5sum: f75362ab9cd70d96621b3690e952085f
     - path: output/meryl/versions.yml
-      md5sum: 944b7ea81a82bd20174c5042857003fc

From d816f2bc5d5338526919d7e1eea42ca6143513fb Mon Sep 17 00:00:00 2001
From: Chris Cheshire <chris.j.cheshire@gmail.com>
Date: Fri, 13 May 2022 12:01:16 +0100
Subject: [PATCH 06/16] Reactivated all tests

---
 tests/modules/bowtie2/align/test.yml | 70 +++++++++++++---------------
 1 file changed, 33 insertions(+), 37 deletions(-)

diff --git a/tests/modules/bowtie2/align/test.yml b/tests/modules/bowtie2/align/test.yml
index 8a33c928..56a8bf53 100644
--- a/tests/modules/bowtie2/align/test.yml
+++ b/tests/modules/bowtie2/align/test.yml
@@ -5,43 +5,39 @@
     - bowtie2/align
   files:
     - path: ./output/bowtie2/test.bam
-    # - path: output/bowtie2/test.bowtie2.log
-    #   md5sum: 7b8a9e61b7646da1089b041333c41a87
-    # - path: output/bowtie2/versions.yml
-    #   md5sum: 24621c58884fe90c2255ccd1fe4352ae
+    - path: ./output/bowtie2/test.bowtie2.log
+    - path: ./output/bowtie2/versions.yml
+      md5sum: 24621c58884fe90c2255ccd1fe4352ae
 
-# - name: bowtie2 align test_bowtie2_align_paired_end
-#   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config
-#   tags:
-#     - bowtie2
-#     - bowtie2/align
-#   files:
-#     - path: output/bowtie2/test.bam
-#     - path: output/bowtie2/test.bowtie2.log
-#       md5sum: bd89ce1b28c93bf822bae391ffcedd19
-#     - path: output/bowtie2/versions.yml
-#       md5sum: f9712ca6d75393ad5c1781a2dcf82d32
+- name: bowtie2 align test_bowtie2_align_paired_end
+  command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config
+  tags:
+    - bowtie2
+    - bowtie2/align
+  files:
+    - path: ./output/bowtie2/test.bam
+    - path: ./output/bowtie2/test.bowtie2.log
+    - path: ./output/bowtie2/versions.yml
+      md5sum: f9712ca6d75393ad5c1781a2dcf82d32
 
-# - name: bowtie2 align test_bowtie2_align_single_end_large_index
-#   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_single_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
-#   tags:
-#     - bowtie2
-#     - bowtie2/align
-#   files:
-#     - path: output/bowtie2/test.bam
-#     - path: output/bowtie2/test.bowtie2.log
-#       md5sum: 7b8a9e61b7646da1089b041333c41a87
-#     - path: output/bowtie2/versions.yml
-#       md5sum: bf0537964a85ce9461ae1b0e2f260211
+- name: bowtie2 align test_bowtie2_align_single_end_large_index
+  command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_single_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
+  tags:
+    - bowtie2
+    - bowtie2/align
+  files:
+    - path: ./output/bowtie2/test.bam
+    - path: ./output/bowtie2/test.bowtie2.log
+    - path: ./output/bowtie2/versions.yml
+      md5sum: bf0537964a85ce9461ae1b0e2f260211
 
-# - name: bowtie2 align test_bowtie2_align_paired_end_large_index
-#   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
-#   tags:
-#     - bowtie2
-#     - bowtie2/align
-#   files:
-#     - path: output/bowtie2/test.bam
-#     - path: output/bowtie2/test.bowtie2.log
-#       md5sum: bd89ce1b28c93bf822bae391ffcedd19
-#     - path: output/bowtie2/versions.yml
-#       md5sum: c0a88953500eaf46de7ff378a7de4a37
+- name: bowtie2 align test_bowtie2_align_paired_end_large_index
+  command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
+  tags:
+    - bowtie2
+    - bowtie2/align
+  files:
+    - path: ./output/bowtie2/test.bam
+    - path: ./output/bowtie2/test.bowtie2.log
+    - path: ./output/bowtie2/versions.yml
+      md5sum: c0a88953500eaf46de7ff378a7de4a37

From ea55129ba3e3cef1a38df236c95d3d4daca63265 Mon Sep 17 00:00:00 2001
From: Chris Cheshire <chris.j.cheshire@gmail.com>
Date: Fri, 13 May 2022 12:20:41 +0100
Subject: [PATCH 07/16] Removed version file checksums

---
 tests/modules/bowtie2/align/test.yml | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/tests/modules/bowtie2/align/test.yml b/tests/modules/bowtie2/align/test.yml
index 56a8bf53..103d65e9 100644
--- a/tests/modules/bowtie2/align/test.yml
+++ b/tests/modules/bowtie2/align/test.yml
@@ -7,7 +7,6 @@
     - path: ./output/bowtie2/test.bam
     - path: ./output/bowtie2/test.bowtie2.log
     - path: ./output/bowtie2/versions.yml
-      md5sum: 24621c58884fe90c2255ccd1fe4352ae
 
 - name: bowtie2 align test_bowtie2_align_paired_end
   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config
@@ -18,7 +17,6 @@
     - path: ./output/bowtie2/test.bam
     - path: ./output/bowtie2/test.bowtie2.log
     - path: ./output/bowtie2/versions.yml
-      md5sum: f9712ca6d75393ad5c1781a2dcf82d32
 
 - name: bowtie2 align test_bowtie2_align_single_end_large_index
   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_single_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
@@ -29,7 +27,6 @@
     - path: ./output/bowtie2/test.bam
     - path: ./output/bowtie2/test.bowtie2.log
     - path: ./output/bowtie2/versions.yml
-      md5sum: bf0537964a85ce9461ae1b0e2f260211
 
 - name: bowtie2 align test_bowtie2_align_paired_end_large_index
   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end_large_index -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config --force_large_index
@@ -40,4 +37,3 @@
     - path: ./output/bowtie2/test.bam
     - path: ./output/bowtie2/test.bowtie2.log
     - path: ./output/bowtie2/versions.yml
-      md5sum: c0a88953500eaf46de7ff378a7de4a37

From 8a17b6dd0e247746cee97e7dc6fa0761534f13ec Mon Sep 17 00:00:00 2001
From: Nicolas Vannieuwkerke <nicolas.vannieuwkerke@ugent.be>
Date: Fri, 13 May 2022 13:46:34 +0200
Subject: [PATCH 08/16] added gatk4/splitintervals

---
 modules/gatk4/splitintervals/main.nf          | 47 +++++++++++++++++++
 modules/gatk4/splitintervals/meta.yml         | 41 ++++++++++++++++
 tests/config/pytest_modules.yml               | 12 +++--
 tests/modules/gatk4/splitintervals/main.nf    | 33 +++++++++++++
 .../gatk4/splitintervals/nextflow.config      |  9 ++++
 tests/modules/gatk4/splitintervals/test.yml   | 25 ++++++++++
 6 files changed, 163 insertions(+), 4 deletions(-)
 create mode 100644 modules/gatk4/splitintervals/main.nf
 create mode 100644 modules/gatk4/splitintervals/meta.yml
 create mode 100644 tests/modules/gatk4/splitintervals/main.nf
 create mode 100644 tests/modules/gatk4/splitintervals/nextflow.config
 create mode 100644 tests/modules/gatk4/splitintervals/test.yml

diff --git a/modules/gatk4/splitintervals/main.nf b/modules/gatk4/splitintervals/main.nf
new file mode 100644
index 00000000..d9407b95
--- /dev/null
+++ b/modules/gatk4/splitintervals/main.nf
@@ -0,0 +1,47 @@
+process GATK4_SPLITINTERVALS {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::gatk4=4.2.6.1" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/gatk4:4.2.6.1--hdfd78af_0':
+        'quay.io/biocontainers/gatk4:4.2.6.1--hdfd78af_0' }"
+
+    input:
+    tuple val(meta), path(intervals)
+    path(fasta)
+    path(fasta_fai)
+    path(fasta_dict)
+
+    output:
+    tuple val(meta), path("**.interval_list"), emit: split_intervals
+    path "versions.yml"           , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def reference = fasta ? "--reference $fasta" : ""
+
+    def avail_mem = 3
+    if (!task.memory) {
+        log.info '[GATK IntervalListToBed] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = task.memory.giga
+    }
+
+    """
+    gatk --java-options "-Xmx${avail_mem}g" SplitIntervals \\
+        --output ${prefix} \\
+        --intervals $intervals \\
+        $reference \\
+        $args 
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gatk4: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+}
diff --git a/modules/gatk4/splitintervals/meta.yml b/modules/gatk4/splitintervals/meta.yml
new file mode 100644
index 00000000..6cf7cedc
--- /dev/null
+++ b/modules/gatk4/splitintervals/meta.yml
@@ -0,0 +1,41 @@
+name: gatk4_splitintervals
+keywords:
+  - interval
+  - bed
+tools:
+  - gatk4:
+      description: Genome Analysis Toolkit (GATK4)
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us/categories/360002369672s
+      tool_dev_url: https://github.com/broadinstitute/gatk
+      doi: "10.1158/1538-7445.AM2017-3590"
+      licence: ["BSD-3-clause"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test' ]
+  - interval:
+      type: file
+      description: Interval list or BED
+      pattern: "*.{interval,interval_list,bed}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test' ]
+  - bed:
+      type: file
+      description: A list of scattered interval lists
+      pattern: "*.interval_list"
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+
+authors:
+  - "@nvnieuwk"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 428c3652..bfd8554a 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -819,6 +819,10 @@ gatk4/selectvariants:
   - modules/gatk4/selectvariants/**
   - tests/modules/gatk4/selectvariants/**
 
+gatk4/splitintervals:
+  - modules/gatk4/splitintervals/**
+  - tests/modules/gatk4/splitintervals/**
+
 gatk4/splitncigarreads:
   - modules/gatk4/splitncigarreads/**
   - tests/modules/gatk4/splitncigarreads/**
@@ -1647,14 +1651,14 @@ samtools/bam2fq:
   - modules/samtools/bam2fq/**
   - tests/modules/samtools/bam2fq/**
 
-samtools/convert:
-  - modules/samtools/convert/**
-  - tests/modules/samtools/convert/**
-
 samtools/collatefastq:
   - modules/samtools/collatefastq/**
   - tests/modules/samtools/collatefastq/**
 
+samtools/convert:
+  - modules/samtools/convert/**
+  - tests/modules/samtools/convert/**
+
 samtools/depth:
   - modules/samtools/depth/**
   - tests/modules/samtools/depth/**
diff --git a/tests/modules/gatk4/splitintervals/main.nf b/tests/modules/gatk4/splitintervals/main.nf
new file mode 100644
index 00000000..f507ece5
--- /dev/null
+++ b/tests/modules/gatk4/splitintervals/main.nf
@@ -0,0 +1,33 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_SPLITINTERVALS } from '../../../../modules/gatk4/splitintervals/main.nf'
+
+workflow test_gatk4_splitintervals_bed {
+    
+    input = [
+        [ id:'test' ], // meta map
+        file(params.test_data['homo_sapiens']['genome']['genome_multi_interval_bed'], checkIfExists: true)
+    ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fasta_fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    fasta_dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+
+    GATK4_SPLITINTERVALS ( input, fasta, fasta_fai, fasta_dict)
+}
+
+workflow test_gatk4_splitintervals_intervals {
+    
+    input = [
+        [ id:'test' ], // meta map
+        file(params.test_data['homo_sapiens']['genome']['genome_interval_list'], checkIfExists: true)
+    ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fasta_fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    fasta_dict = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+
+    GATK4_SPLITINTERVALS ( input, fasta, fasta_fai, fasta_dict)
+}
\ No newline at end of file
diff --git a/tests/modules/gatk4/splitintervals/nextflow.config b/tests/modules/gatk4/splitintervals/nextflow.config
new file mode 100644
index 00000000..10fda96c
--- /dev/null
+++ b/tests/modules/gatk4/splitintervals/nextflow.config
@@ -0,0 +1,9 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+
+    withName: GATK4_SPLITINTERVALS {
+        ext.args = "--scatter-count 2"
+    }
+}
\ No newline at end of file
diff --git a/tests/modules/gatk4/splitintervals/test.yml b/tests/modules/gatk4/splitintervals/test.yml
new file mode 100644
index 00000000..dc0ea6ba
--- /dev/null
+++ b/tests/modules/gatk4/splitintervals/test.yml
@@ -0,0 +1,25 @@
+- name: gatk4 splitintervals test_gatk4_splitintervals_bed
+  command: nextflow run tests/modules/gatk4/splitintervals -entry test_gatk4_splitintervals_bed -c tests/config/nextflow.config
+  tags:
+    - gatk4/splitintervals
+    - gatk4
+  files:
+    - path: output/gatk4/test/0000-scattered.interval_list
+      md5sum: c8d6b19e7a92535b6ce9608eae558faa
+    - path: output/gatk4/test/0001-scattered.interval_list
+      md5sum: b1877ad96aec308906594c50ebbe3ded
+    - path: output/gatk4/versions.yml
+      md5sum: c7b9068d84e0e05412f50e86d5b87078
+
+- name: gatk4 splitintervals test_gatk4_splitintervals_intervals
+  command: nextflow run tests/modules/gatk4/splitintervals -entry test_gatk4_splitintervals_intervals -c tests/config/nextflow.config
+  tags:
+    - gatk4/splitintervals
+    - gatk4
+  files:
+    - path: output/gatk4/test/0000-scattered.interval_list
+      md5sum: ebd6b34a335efc6732ff541936c6d2d5
+    - path: output/gatk4/test/0001-scattered.interval_list
+      md5sum: 9459b0e124fa84ec1e64ac4615bc9af7
+    - path: output/gatk4/versions.yml
+      md5sum: 4eec8245944ad3be4cc22403c6ffb877

From 9d53577282ecb9d683ba93f9741dd4e721b2fcba Mon Sep 17 00:00:00 2001
From: Nicolas Vannieuwkerke <nicolas.vannieuwkerke@ugent.be>
Date: Fri, 13 May 2022 13:48:33 +0200
Subject: [PATCH 09/16] linting

---
 modules/gatk4/splitintervals/main.nf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules/gatk4/splitintervals/main.nf b/modules/gatk4/splitintervals/main.nf
index d9407b95..d575191f 100644
--- a/modules/gatk4/splitintervals/main.nf
+++ b/modules/gatk4/splitintervals/main.nf
@@ -37,7 +37,7 @@ process GATK4_SPLITINTERVALS {
         --output ${prefix} \\
         --intervals $intervals \\
         $reference \\
-        $args 
+        $args
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":

From d1f824ae66f432301bf1325ab8fa48825aa43bac Mon Sep 17 00:00:00 2001
From: Nicolas Vannieuwkerke <nicolas.vannieuwkerke@ugent.be>
Date: Fri, 13 May 2022 14:21:48 +0200
Subject: [PATCH 10/16] updated meta.yml

---
 modules/gatk4/splitintervals/main.nf  |  2 +-
 modules/gatk4/splitintervals/meta.yml | 12 ++++++++++++
 2 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/modules/gatk4/splitintervals/main.nf b/modules/gatk4/splitintervals/main.nf
index d575191f..adb72dfb 100644
--- a/modules/gatk4/splitintervals/main.nf
+++ b/modules/gatk4/splitintervals/main.nf
@@ -11,7 +11,7 @@ process GATK4_SPLITINTERVALS {
     tuple val(meta), path(intervals)
     path(fasta)
     path(fasta_fai)
-    path(fasta_dict)
+    path(dict)
 
     output:
     tuple val(meta), path("**.interval_list"), emit: split_intervals
diff --git a/modules/gatk4/splitintervals/meta.yml b/modules/gatk4/splitintervals/meta.yml
index 6cf7cedc..ba557544 100644
--- a/modules/gatk4/splitintervals/meta.yml
+++ b/modules/gatk4/splitintervals/meta.yml
@@ -21,6 +21,18 @@ input:
       type: file
       description: Interval list or BED
       pattern: "*.{interval,interval_list,bed}"
+  - fasta:
+      type: file
+      description: Reference FASTA
+      pattern: "*.{fa,fasta}"
+  - fasta_fai:
+      type: file
+      description: Reference FASTA index
+      pattern: "*.fai"
+  - dict:
+      type: file
+      description: Reference sequence dictionary
+      pattern: "*.dict"
 
 output:
   - meta:

From 1c2a99142928c6ba2d32226168e6755263784474 Mon Sep 17 00:00:00 2001
From: Nicolas Vannieuwkerke <nicolas.vannieuwkerke@ugent.be>
Date: Fri, 13 May 2022 14:25:53 +0200
Subject: [PATCH 11/16] fixed memory warning

---
 modules/gatk4/splitintervals/main.nf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules/gatk4/splitintervals/main.nf b/modules/gatk4/splitintervals/main.nf
index adb72dfb..3ab5a71b 100644
--- a/modules/gatk4/splitintervals/main.nf
+++ b/modules/gatk4/splitintervals/main.nf
@@ -27,7 +27,7 @@ process GATK4_SPLITINTERVALS {
 
     def avail_mem = 3
     if (!task.memory) {
-        log.info '[GATK IntervalListToBed] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+        log.info '[GATK SplitIntervals] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
     } else {
         avail_mem = task.memory.giga
     }

From a72f06dfe8ce6f29d90f3f952bf5700019ff4d14 Mon Sep 17 00:00:00 2001
From: Nicolas Vannieuwkerke <nicolas.vannieuwkerke@ugent.be>
Date: Fri, 13 May 2022 14:27:59 +0200
Subject: [PATCH 12/16] removed versions checksum

---
 tests/modules/gatk4/splitintervals/test.yml | 2 --
 1 file changed, 2 deletions(-)

diff --git a/tests/modules/gatk4/splitintervals/test.yml b/tests/modules/gatk4/splitintervals/test.yml
index dc0ea6ba..741c6bec 100644
--- a/tests/modules/gatk4/splitintervals/test.yml
+++ b/tests/modules/gatk4/splitintervals/test.yml
@@ -9,7 +9,6 @@
     - path: output/gatk4/test/0001-scattered.interval_list
       md5sum: b1877ad96aec308906594c50ebbe3ded
     - path: output/gatk4/versions.yml
-      md5sum: c7b9068d84e0e05412f50e86d5b87078
 
 - name: gatk4 splitintervals test_gatk4_splitintervals_intervals
   command: nextflow run tests/modules/gatk4/splitintervals -entry test_gatk4_splitintervals_intervals -c tests/config/nextflow.config
@@ -22,4 +21,3 @@
     - path: output/gatk4/test/0001-scattered.interval_list
       md5sum: 9459b0e124fa84ec1e64ac4615bc9af7
     - path: output/gatk4/versions.yml
-      md5sum: 4eec8245944ad3be4cc22403c6ffb877

From 6427eeeca04790574303b13ba473594fdf725821 Mon Sep 17 00:00:00 2001
From: Nicolas Vannieuwkerke <nicolas.vannieuwkerke@ugent.be>
Date: Fri, 13 May 2022 14:35:22 +0200
Subject: [PATCH 13/16] added tmp-dir + alignment correction

---
 modules/gatk4/splitintervals/main.nf | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/modules/gatk4/splitintervals/main.nf b/modules/gatk4/splitintervals/main.nf
index 3ab5a71b..9df66c21 100644
--- a/modules/gatk4/splitintervals/main.nf
+++ b/modules/gatk4/splitintervals/main.nf
@@ -15,7 +15,7 @@ process GATK4_SPLITINTERVALS {
 
     output:
     tuple val(meta), path("**.interval_list"), emit: split_intervals
-    path "versions.yml"           , emit: versions
+    path "versions.yml"                      , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
@@ -37,6 +37,7 @@ process GATK4_SPLITINTERVALS {
         --output ${prefix} \\
         --intervals $intervals \\
         $reference \\
+        --tmp-dir . \\
         $args
 
     cat <<-END_VERSIONS > versions.yml

From cbc47767f7bb4751a9d178395bd66f051401a0c1 Mon Sep 17 00:00:00 2001
From: Matthias De Smet <11850640+matthdsm@users.noreply.github.com>
Date: Fri, 13 May 2022 15:26:41 +0200
Subject: [PATCH 14/16] add sort option to bowtie2

---
 modules/bowtie2/align/main.nf        |  6 +++---
 modules/bowtie2/align/meta.yml       |  9 +++++++++
 tests/modules/bowtie2/align/main.nf  | 29 +++++++++++++++++++++++-----
 tests/modules/bowtie2/align/test.yml | 10 ++++++++++
 4 files changed, 46 insertions(+), 8 deletions(-)

diff --git a/modules/bowtie2/align/main.nf b/modules/bowtie2/align/main.nf
index e4bb4327..c74e376f 100644
--- a/modules/bowtie2/align/main.nf
+++ b/modules/bowtie2/align/main.nf
@@ -11,6 +11,7 @@ process BOWTIE2_ALIGN {
     tuple val(meta), path(reads)
     path  index
     val   save_unaligned
+    val   sort_bam
 
     output:
     tuple val(meta), path("*.bam")    , emit: bam
@@ -36,8 +37,7 @@ process BOWTIE2_ALIGN {
         reads_args = "-1 ${reads[0]} -2 ${reads[1]}"
     }
 
-    def samtools_command = "samtools view -@ $task.cpus --bam --with-header ${args2} > ${prefix}.bam"
-
+    def samtools_command = sort_bam ? 'sort' : 'view'
 
     """
     INDEX=`find -L ./ -name "*.rev.1.bt2" | sed "s/.rev.1.bt2//"`
@@ -51,7 +51,7 @@ process BOWTIE2_ALIGN {
         $unaligned \\
         $args \\
         2> ${prefix}.bowtie2.log \\
-        | $samtools_command
+        | samtools $samtools_command $args2 --threads $task.cpus -o ${prefix}.bam -
 
     if [ -f ${prefix}.unmapped.fastq.1.gz ]; then
         mv ${prefix}.unmapped.fastq.1.gz ${prefix}.unmapped_1.fastq.gz
diff --git a/modules/bowtie2/align/meta.yml b/modules/bowtie2/align/meta.yml
index f80421ec..c99fa4e3 100644
--- a/modules/bowtie2/align/meta.yml
+++ b/modules/bowtie2/align/meta.yml
@@ -29,6 +29,15 @@ input:
       type: file
       description: Bowtie2 genome index files
       pattern: "*.ebwt"
+  - save_unaligned:
+      type: boolean
+      description: |
+        Save reads that do not map to the reference (true) or discard them (false)
+        (default: false)
+  - sort_bam:
+      type: boolean
+      description: use samtools sort (true) or samtools view (false)
+      pattern: "true or false"
 output:
   - bam:
       type: file
diff --git a/tests/modules/bowtie2/align/main.nf b/tests/modules/bowtie2/align/main.nf
index 42e2306a..75ab33ac 100644
--- a/tests/modules/bowtie2/align/main.nf
+++ b/tests/modules/bowtie2/align/main.nf
@@ -14,9 +14,25 @@ workflow test_bowtie2_align_single_end {
     ]
     fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
     save_unaligned = false
+    sort = false
 
     BOWTIE2_BUILD ( fasta )
-    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned )
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned, sort )
+}
+
+workflow test_bowtie2_align_single_end_sorted {
+    input = [
+        [ id:'test', single_end:true ], // meta map
+        [
+            file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+        ]
+    ]
+    fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    save_unaligned = false
+    sort = true
+
+    BOWTIE2_BUILD ( fasta )
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned, sort )
 }
 
 workflow test_bowtie2_align_paired_end {
@@ -29,9 +45,10 @@ workflow test_bowtie2_align_paired_end {
     ]
     fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
     save_unaligned = false
+    sort = false
 
     BOWTIE2_BUILD ( fasta )
-    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned )
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned, sort )
 }
 
 workflow test_bowtie2_align_single_end_large_index {
@@ -43,9 +60,10 @@ workflow test_bowtie2_align_single_end_large_index {
     ]
     fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
     save_unaligned = false
+    sort = false
 
     BOWTIE2_BUILD ( fasta )
-    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned )
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned, sort )
 }
 
 workflow test_bowtie2_align_paired_end_large_index {
@@ -58,7 +76,8 @@ workflow test_bowtie2_align_paired_end_large_index {
     ]
     fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
     save_unaligned = false
+    sort = false
 
     BOWTIE2_BUILD ( fasta )
-    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned )
-}
\ No newline at end of file
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned, sort )
+}
diff --git a/tests/modules/bowtie2/align/test.yml b/tests/modules/bowtie2/align/test.yml
index 103d65e9..4f8e4ff1 100644
--- a/tests/modules/bowtie2/align/test.yml
+++ b/tests/modules/bowtie2/align/test.yml
@@ -8,6 +8,16 @@
     - path: ./output/bowtie2/test.bowtie2.log
     - path: ./output/bowtie2/versions.yml
 
+- name: bowtie2 align test_bowtie2_align_single_end
+  command: nextflow run ./tests/modules/bowtie2/align -entry test_bowtie2_align_single_end_sorted -c ./tests/config/nextflow.config -c ./tests/modules/bowtie2/align/nextflow.config
+  tags:
+    - bowtie2
+    - bowtie2/align
+  files:
+    - path: ./output/bowtie2/test.bam
+    - path: ./output/bowtie2/test.bowtie2.log
+    - path: ./output/bowtie2/versions.yml
+
 - name: bowtie2 align test_bowtie2_align_paired_end
   command: nextflow run tests/modules/bowtie2/align -entry test_bowtie2_align_paired_end -c tests/config/nextflow.config -c tests/modules/bowtie2/align/nextflow.config
   tags:

From 04f82d4d349573813964ff74ddddc79a4323e88a Mon Sep 17 00:00:00 2001
From: Matthias De Smet <11850640+matthdsm@users.noreply.github.com>
Date: Fri, 13 May 2022 15:32:14 +0200
Subject: [PATCH 15/16] fix test name

---
 tests/modules/bowtie2/align/test.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/modules/bowtie2/align/test.yml b/tests/modules/bowtie2/align/test.yml
index 4f8e4ff1..9c9ff7a1 100644
--- a/tests/modules/bowtie2/align/test.yml
+++ b/tests/modules/bowtie2/align/test.yml
@@ -8,7 +8,7 @@
     - path: ./output/bowtie2/test.bowtie2.log
     - path: ./output/bowtie2/versions.yml
 
-- name: bowtie2 align test_bowtie2_align_single_end
+- name: bowtie2 align test_bowtie2_align_single_end_sorted
   command: nextflow run ./tests/modules/bowtie2/align -entry test_bowtie2_align_single_end_sorted -c ./tests/config/nextflow.config -c ./tests/modules/bowtie2/align/nextflow.config
   tags:
     - bowtie2

From 5a2e3035220f30ab2d6fec5ded5a365e327248e4 Mon Sep 17 00:00:00 2001
From: Matthias De Smet <11850640+matthdsm@users.noreply.github.com>
Date: Fri, 13 May 2022 19:16:50 +0200
Subject: [PATCH 16/16] add PE sorted test

---
 tests/modules/bowtie2/align/main.nf | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/tests/modules/bowtie2/align/main.nf b/tests/modules/bowtie2/align/main.nf
index 75ab33ac..4f08b533 100644
--- a/tests/modules/bowtie2/align/main.nf
+++ b/tests/modules/bowtie2/align/main.nf
@@ -51,6 +51,22 @@ workflow test_bowtie2_align_paired_end {
     BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned, sort )
 }
 
+workflow test_bowtie2_align_paired_end_sorted {
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        [
+            file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true),
+            file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true)
+        ]
+    ]
+    fasta = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    save_unaligned = false
+    sort = true
+
+    BOWTIE2_BUILD ( fasta )
+    BOWTIE2_ALIGN ( input, BOWTIE2_BUILD.out.index, save_unaligned, sort )
+}
+
 workflow test_bowtie2_align_single_end_large_index {
     input = [
         [ id:'test', single_end:true ], // meta map