From 064712c4ea2b42a405a98e55a4aac81c3f79d3ff Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 09:11:45 +0200
Subject: [PATCH 01/12] added cram_bam conversion & tumor_only check

---
 modules/cnvkit/batch/main.nf | 33 ++++++++++++++++++++++++++++-----
 1 file changed, 28 insertions(+), 5 deletions(-)

diff --git a/modules/cnvkit/batch/main.nf b/modules/cnvkit/batch/main.nf
index 7c44d9f6..1ceb7072 100644
--- a/modules/cnvkit/batch/main.nf
+++ b/modules/cnvkit/batch/main.nf
@@ -5,7 +5,7 @@ process CNVKIT_BATCH {
     conda (params.enable_conda ? 'bioconda::cnvkit=0.9.9' : null)
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
         'https://depot.galaxyproject.org/singularity/cnvkit:0.9.9--pyhdfd78af_0' :
-        'quay.io/biocontainers/cnvkit:0.9.9--pyhdfd78af_0' }"
+        'quay.io/biocontainers/mulled-v2-780d630a9bb6a0ff2e7b6f730906fd703e40e98f:304d1c5ab610f216e77c61420ebe85f1e7c5968a-0' }"
 
     input:
     tuple val(meta), path(tumor), path(normal)
@@ -18,6 +18,8 @@ process CNVKIT_BATCH {
     tuple val(meta), path("*.cnn"), emit: cnn, optional: true
     tuple val(meta), path("*.cnr"), emit: cnr, optional: true
     tuple val(meta), path("*.cns"), emit: cns, optional: true
+    tuple val(meta), path("*.pdf"), emit: pdf, optional: true
+    tuple val(meta), path("*.png"), emit: png, optional: true
     path "versions.yml"           , emit: versions
 
     when:
@@ -25,21 +27,42 @@ process CNVKIT_BATCH {
 
     script:
     def args = task.ext.args ?: ''
-    def normal_args = normal ? "--normal $normal" : ""
-    def fasta_args = fasta ? "--fasta $fasta" : ""
+    // execute samtools only when cram files are input
+    // input pair is assumed to have same extension if both exist
+    def is_cram = tumor.Extension == "cram" ? true : false
+    def tumor_out = is_cram ? tumor.BaseName + ".bam" : "${tumor}"
+    // do not run samtools on normal samples in tumor_only
+    def normal_exists = normal ? true: false
+    // tumor_only mode does not need fasta & target
+    // instead it requires a pre-computed reference which is built from fasta & target
+    def (normal_out, normal_args, fasta_args, target_args) = ["", "", "", ""]
     def reference_args = reference ? "--reference $reference" : ""
 
-    def target_args = ""
+    if (normal_exists){
+        def normal_prefix = normal.BaseName
+        normal_out = is_cram ? "${normal_prefix}" + ".bam" : "${normal}"
+        normal_args = normal_prefix ? "--normal $normal_out" : ""
+        fasta_args = fasta ? "--fasta $fasta" : ""
+    }
+
     if (args.contains("--method wgs") || args.contains("-m wgs")) {
         target_args = targets ? "--targets $targets" : ""
     }
     else {
         target_args = "--targets $targets"
     }
+
     """
+    if $is_cram; then
+        samtools view -T $fasta $tumor -@ $task.cpus -o $tumor_out
+        if $normal_exists; then
+            samtools view -T $fasta $normal -@ $task.cpus -o $normal_out
+        fi
+    fi
+
     cnvkit.py \\
         batch \\
-        $tumor \\
+        $tumor_out \\
         $normal_args \\
         $fasta_args \\
         $reference_args \\

From 5a1297cc7d55de78cdb945977f8e963fe5ecde78 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 09:12:47 +0200
Subject: [PATCH 02/12] added optional pdf/png outputs

---
 modules/cnvkit/batch/meta.yml | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/modules/cnvkit/batch/meta.yml b/modules/cnvkit/batch/meta.yml
index 474c55f2..03277e33 100644
--- a/modules/cnvkit/batch/meta.yml
+++ b/modules/cnvkit/batch/meta.yml
@@ -80,6 +80,14 @@ output:
       type: file
       description: File containing copy number segment information
       pattern: "*.{cns}"
+  - pdf:
+      type: file
+      description: File with plot of copy numbers or segments on chromosomes
+      pattern: "*.{pdf}"
+  - png:
+      type: file
+      description: File with plot of bin-level log2 coverages and segmentation calls
+      pattern: "*.{png}"
   - versions:
       type: file
       description: File containing software versions
@@ -91,3 +99,4 @@ authors:
   - "@drpatelh"
   - "@fbdtemme"
   - "@lassefolkersen"
+  - "@SusiJo"

From 3d57caa683a48a014abf986cfdaaf0b9f6ff9676 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 09:13:27 +0200
Subject: [PATCH 03/12] added test & corrected cram input

---
 tests/modules/cnvkit/batch/main.nf  | 19 ++++++++++++++---
 tests/modules/cnvkit/batch/test.yml | 33 +++++++++++++++++++----------
 2 files changed, 38 insertions(+), 14 deletions(-)

diff --git a/tests/modules/cnvkit/batch/main.nf b/tests/modules/cnvkit/batch/main.nf
index 6b40dec6..15b5dfe6 100755
--- a/tests/modules/cnvkit/batch/main.nf
+++ b/tests/modules/cnvkit/batch/main.nf
@@ -35,8 +35,8 @@ workflow test_cnvkit_cram {
 
     input = [
         [ id:'test'], // meta map
-        file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam'], checkIfExists: true),
-        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true)
+        file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_cram'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true)
     ]
     fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
 
@@ -53,5 +53,18 @@ workflow test_cnvkit_tumoronly {
     fasta     = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
     reference = file(params.test_data['generic']['cnn']['reference'], checkIfExists: true)
 
-    CNVKIT_TUMORONLY ( input, [], [], reference )
+    CNVKIT_TUMORONLY ( input, fasta, [], reference )
+}
+
+workflow test_cnvkit_tumoronly_cram {
+
+    input = [
+        [ id:'test'], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_cram'], checkIfExists: true),
+        []
+    ]
+    fasta     = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    reference = file(params.test_data['generic']['cnn']['reference'], checkIfExists: true)
+
+    CNVKIT_TUMORONLY ( input, fasta, [], reference )
 }
diff --git a/tests/modules/cnvkit/batch/test.yml b/tests/modules/cnvkit/batch/test.yml
index 57af3603..9bce265a 100755
--- a/tests/modules/cnvkit/batch/test.yml
+++ b/tests/modules/cnvkit/batch/test.yml
@@ -1,15 +1,15 @@
 - name: cnvkit batch test_cnvkit_hybrid
-  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_hybrid -c ./tests/config/nextflow.config -c ./tests/modules/cnvkit/batch/nextflow.config
+  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_hybrid -c ./tests/config/nextflow.config  -c ./tests/modules/cnvkit/batch/nextflow.config
   tags:
-    - cnvkit/batch
     - cnvkit
+    - cnvkit/batch
   files:
     - path: output/cnvkit/baits.antitarget.bed
       md5sum: d41d8cd98f00b204e9800998ecf8427e
     - path: output/cnvkit/baits.target.bed
       md5sum: 26d25ff2d6c45b6d92169b3559c6acdb
     - path: output/cnvkit/reference.cnn
-      md5sum: ac99c1ad8b917b96ae15119146c91ab9
+      md5sum: 035d031f54c5f1b43b903da96559b475
     - path: output/cnvkit/test.paired_end.sorted.antitargetcoverage.cnn
       md5sum: 203caf8cef6935bb50b4138097955cb8
     - path: output/cnvkit/test.paired_end.sorted.bintest.cns
@@ -28,10 +28,10 @@
       md5sum: aa8a018b1d4d1e688c9f9f6ae01bf4d7
 
 - name: cnvkit batch test_cnvkit_wgs
-  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_wgs -c ./tests/config/nextflow.config -c ./tests/modules/cnvkit/batch/nextflow.config
+  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_wgs -c ./tests/config/nextflow.config  -c ./tests/modules/cnvkit/batch/nextflow.config
   tags:
-    - cnvkit/batch
     - cnvkit
+    - cnvkit/batch
   files:
     - path: output/cnvkit/genome.antitarget.bed
       md5sum: d41d8cd98f00b204e9800998ecf8427e
@@ -40,7 +40,7 @@
     - path: output/cnvkit/genome.target.bed
       md5sum: a13353ae9c8405e701390c069255bbd2
     - path: output/cnvkit/reference.cnn
-      md5sum: 05c6211e0179885b8a83e44fd21d5f86
+      md5sum: 1606a85410bfaa79464be6e98699aa83
     - path: output/cnvkit/test.paired_end.sorted.antitargetcoverage.cnn
       md5sum: 203caf8cef6935bb50b4138097955cb8
     - path: output/cnvkit/test.paired_end.sorted.targetcoverage.cnn
@@ -59,10 +59,10 @@
       md5sum: 6ae6b3fce7299eedca6133d911c38fe1
 
 - name: cnvkit batch test_cnvkit_cram
-  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_cram -c ./tests/config/nextflow.config -c ./tests/modules/cnvkit/batch/nextflow.config
+  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_cram -c ./tests/config/nextflow.config  -c ./tests/modules/cnvkit/batch/nextflow.config
   tags:
-    - cnvkit/batch
     - cnvkit
+    - cnvkit/batch
   files:
     - path: output/cnvkit/genome.antitarget.bed
       md5sum: d41d8cd98f00b204e9800998ecf8427e
@@ -71,7 +71,7 @@
     - path: output/cnvkit/genome.target.bed
       md5sum: a13353ae9c8405e701390c069255bbd2
     - path: output/cnvkit/reference.cnn
-      md5sum: 05c6211e0179885b8a83e44fd21d5f86
+      md5sum: 1606a85410bfaa79464be6e98699aa83
     - path: output/cnvkit/test.paired_end.sorted.antitargetcoverage.cnn
       md5sum: 203caf8cef6935bb50b4138097955cb8
     - path: output/cnvkit/test.paired_end.sorted.targetcoverage.cnn
@@ -90,10 +90,21 @@
       md5sum: 6ae6b3fce7299eedca6133d911c38fe1
 
 - name: cnvkit batch test_cnvkit_tumoronly
-  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_tumoronly -c ./tests/config/nextflow.config -c ./tests/modules/cnvkit/batch/nextflow.config
+  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_tumoronly -c ./tests/config/nextflow.config  -c ./tests/modules/cnvkit/batch/nextflow.config
   tags:
-    - cnvkit/batch
     - cnvkit
+    - cnvkit/batch
+  files:
+    - path: output/cnvkit/reference.antitarget-tmp.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/cnvkit/reference.target-tmp.bed
+      md5sum: 26d25ff2d6c45b6d92169b3559c6acdb
+
+- name: cnvkit batch test_cnvkit_tumoronly_cram
+  command: nextflow run ./tests/modules/cnvkit/batch -entry test_cnvkit_tumoronly_cram -c ./tests/config/nextflow.config  -c ./tests/modules/cnvkit/batch/nextflow.config
+  tags:
+    - cnvkit
+    - cnvkit/batch
   files:
     - path: output/cnvkit/reference.antitarget-tmp.bed
       md5sum: d41d8cd98f00b204e9800998ecf8427e

From 129270d7bc31830ea5f4772264c6dcb5007ee817 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 10:37:46 +0200
Subject: [PATCH 04/12] adjusted conda env & singularity container

---
 modules/cnvkit/batch/main.nf | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/modules/cnvkit/batch/main.nf b/modules/cnvkit/batch/main.nf
index 1ceb7072..063808a1 100644
--- a/modules/cnvkit/batch/main.nf
+++ b/modules/cnvkit/batch/main.nf
@@ -2,9 +2,9 @@ process CNVKIT_BATCH {
     tag "$meta.id"
     label 'process_low'
 
-    conda (params.enable_conda ? 'bioconda::cnvkit=0.9.9' : null)
+    conda (params.enable_conda ? 'bioconda::cnvkit=0.9.9 bioconda::samtools=1.15.1' : null)
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/cnvkit:0.9.9--pyhdfd78af_0' :
+        'https://depot.galaxyproject.org/singularity/mulled-v2-780d630a9bb6a0ff2e7b6f730906fd703e40e98f:304d1c5ab610f216e77c61420ebe85f1e7c5968a-0' :
         'quay.io/biocontainers/mulled-v2-780d630a9bb6a0ff2e7b6f730906fd703e40e98f:304d1c5ab610f216e77c61420ebe85f1e7c5968a-0' }"
 
     input:

From f392c4bed533ccacd0e79cca736666417e877ad0 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 13:05:43 +0200
Subject: [PATCH 05/12] removed params section & added description fasta

---
 modules/cnvkit/batch/meta.yml | 23 +----------------------
 1 file changed, 1 insertion(+), 22 deletions(-)

diff --git a/modules/cnvkit/batch/meta.yml b/modules/cnvkit/batch/meta.yml
index 03277e33..2cd675c7 100644
--- a/modules/cnvkit/batch/meta.yml
+++ b/modules/cnvkit/batch/meta.yml
@@ -11,27 +11,6 @@ tools:
       homepage: https://cnvkit.readthedocs.io/en/stable/index.html
       documentation: https://cnvkit.readthedocs.io/en/stable/index.html
       licence: ["Apache-2.0"]
-params:
-  - outdir:
-      type: string
-      description: |
-        The pipeline's output directory. By default, the module will
-        output files into `$params.outdir/<SOFTWARE>`
-  - publish_dir_mode:
-      type: string
-      description: |
-        Value for the Nextflow `publishDir` mode parameter.
-        Available: symlink, rellink, link, copy, copyNoFollow, move.
-  - enable_conda:
-      type: boolean
-      description: |
-        Run the module with Conda using the software specified
-        via the `conda` directive
-  - singularity_pull_docker_container:
-      type: boolean
-      description: |
-        Instead of directly downloading Singularity images for use with Singularity,
-        force the workflow to pull and convert Docker containers instead.
 input:
   - meta:
       type: map
@@ -49,7 +28,7 @@ input:
   - fasta:
       type: file
       description: |
-        Input reference genome fasta file
+        Input reference genome fasta file (only needed for cram_input and/or when normal_samples are provided)
   - targetfile:
       type: file
       description: |

From 00077c36c8f57474c7540053b7ec43d4c4378c13 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 13:06:16 +0200
Subject: [PATCH 06/12] removed unnecessary fasta for tumor_only_bam

---
 tests/modules/cnvkit/batch/main.nf | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/tests/modules/cnvkit/batch/main.nf b/tests/modules/cnvkit/batch/main.nf
index 15b5dfe6..988ed3ad 100755
--- a/tests/modules/cnvkit/batch/main.nf
+++ b/tests/modules/cnvkit/batch/main.nf
@@ -50,10 +50,9 @@ workflow test_cnvkit_tumoronly {
         file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam'], checkIfExists: true),
         []
     ]
-    fasta     = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
     reference = file(params.test_data['generic']['cnn']['reference'], checkIfExists: true)
 
-    CNVKIT_TUMORONLY ( input, fasta, [], reference )
+    CNVKIT_TUMORONLY ( input, [], [], reference )
 }
 
 workflow test_cnvkit_tumoronly_cram {

From 3b3921cfa319602ed15c7021969122d72d600e46 Mon Sep 17 00:00:00 2001
From: SusiJo <43847534+SusiJo@users.noreply.github.com>
Date: Fri, 20 May 2022 13:07:57 +0200
Subject: [PATCH 07/12] refine comment

Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de>
---
 modules/cnvkit/batch/main.nf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules/cnvkit/batch/main.nf b/modules/cnvkit/batch/main.nf
index 063808a1..63482c6c 100644
--- a/modules/cnvkit/batch/main.nf
+++ b/modules/cnvkit/batch/main.nf
@@ -27,7 +27,7 @@ process CNVKIT_BATCH {
 
     script:
     def args = task.ext.args ?: ''
-    // execute samtools only when cram files are input
+    // execute samtools only when cram files are input, cnvkit runs natively on bam but is prohibitively slow 
     // input pair is assumed to have same extension if both exist
     def is_cram = tumor.Extension == "cram" ? true : false
     def tumor_out = is_cram ? tumor.BaseName + ".bam" : "${tumor}"

From 2e5aa0dbb45632072cbe862985e0e4ff50c55867 Mon Sep 17 00:00:00 2001
From: SusiJo <43847534+SusiJo@users.noreply.github.com>
Date: Fri, 20 May 2022 13:08:26 +0200
Subject: [PATCH 08/12] refine comment

Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de>
---
 modules/cnvkit/batch/main.nf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules/cnvkit/batch/main.nf b/modules/cnvkit/batch/main.nf
index 63482c6c..55e3ee6b 100644
--- a/modules/cnvkit/batch/main.nf
+++ b/modules/cnvkit/batch/main.nf
@@ -31,7 +31,7 @@ process CNVKIT_BATCH {
     // input pair is assumed to have same extension if both exist
     def is_cram = tumor.Extension == "cram" ? true : false
     def tumor_out = is_cram ? tumor.BaseName + ".bam" : "${tumor}"
-    // do not run samtools on normal samples in tumor_only
+    // do not run samtools on normal samples in tumor_only mode
     def normal_exists = normal ? true: false
     // tumor_only mode does not need fasta & target
     // instead it requires a pre-computed reference which is built from fasta & target

From 153c900d0e8c2e871a984bf40fb92a06ed019ae7 Mon Sep 17 00:00:00 2001
From: SusiJo <43847534+SusiJo@users.noreply.github.com>
Date: Fri, 20 May 2022 13:08:38 +0200
Subject: [PATCH 09/12] refine comment

Co-authored-by: FriederikeHanssen <Friederike.hanssen@qbic.uni-tuebingen.de>
---
 modules/cnvkit/batch/main.nf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules/cnvkit/batch/main.nf b/modules/cnvkit/batch/main.nf
index 55e3ee6b..c3a4ba32 100644
--- a/modules/cnvkit/batch/main.nf
+++ b/modules/cnvkit/batch/main.nf
@@ -34,7 +34,7 @@ process CNVKIT_BATCH {
     // do not run samtools on normal samples in tumor_only mode
     def normal_exists = normal ? true: false
     // tumor_only mode does not need fasta & target
-    // instead it requires a pre-computed reference which is built from fasta & target
+    // instead it requires a pre-computed reference.cnn which is built from fasta & target
     def (normal_out, normal_args, fasta_args, target_args) = ["", "", "", ""]
     def reference_args = reference ? "--reference $reference" : ""
 

From bdf222649d04fc2748d5247e02aa7fd7a1646a70 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 13:25:57 +0200
Subject: [PATCH 10/12] remove unnecessary if statement

---
 modules/cnvkit/batch/main.nf | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/modules/cnvkit/batch/main.nf b/modules/cnvkit/batch/main.nf
index c3a4ba32..66cfe8b2 100644
--- a/modules/cnvkit/batch/main.nf
+++ b/modules/cnvkit/batch/main.nf
@@ -27,7 +27,7 @@ process CNVKIT_BATCH {
 
     script:
     def args = task.ext.args ?: ''
-    // execute samtools only when cram files are input, cnvkit runs natively on bam but is prohibitively slow 
+    // execute samtools only when cram files are input, cnvkit runs natively on bam but is prohibitively slow
     // input pair is assumed to have same extension if both exist
     def is_cram = tumor.Extension == "cram" ? true : false
     def tumor_out = is_cram ? tumor.BaseName + ".bam" : "${tumor}"
@@ -35,7 +35,8 @@ process CNVKIT_BATCH {
     def normal_exists = normal ? true: false
     // tumor_only mode does not need fasta & target
     // instead it requires a pre-computed reference.cnn which is built from fasta & target
-    def (normal_out, normal_args, fasta_args, target_args) = ["", "", "", ""]
+    def (normal_out, normal_args, fasta_args) = ["", "", ""]
+    def target_args = targets ? "--targets $targets" : ""
     def reference_args = reference ? "--reference $reference" : ""
 
     if (normal_exists){
@@ -45,13 +46,6 @@ process CNVKIT_BATCH {
         fasta_args = fasta ? "--fasta $fasta" : ""
     }
 
-    if (args.contains("--method wgs") || args.contains("-m wgs")) {
-        target_args = targets ? "--targets $targets" : ""
-    }
-    else {
-        target_args = "--targets $targets"
-    }
-
     """
     if $is_cram; then
         samtools view -T $fasta $tumor -@ $task.cpus -o $tumor_out

From 9096be5464d800156761b7c73c574f5a5bad2f72 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Fri, 20 May 2022 13:29:49 +0200
Subject: [PATCH 11/12] improve readability

---
 modules/cnvkit/batch/main.nf | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/modules/cnvkit/batch/main.nf b/modules/cnvkit/batch/main.nf
index 66cfe8b2..c1f9ef87 100644
--- a/modules/cnvkit/batch/main.nf
+++ b/modules/cnvkit/batch/main.nf
@@ -27,17 +27,17 @@ process CNVKIT_BATCH {
 
     script:
     def args = task.ext.args ?: ''
+
     // execute samtools only when cram files are input, cnvkit runs natively on bam but is prohibitively slow
     // input pair is assumed to have same extension if both exist
     def is_cram = tumor.Extension == "cram" ? true : false
     def tumor_out = is_cram ? tumor.BaseName + ".bam" : "${tumor}"
+
     // do not run samtools on normal samples in tumor_only mode
     def normal_exists = normal ? true: false
     // tumor_only mode does not need fasta & target
     // instead it requires a pre-computed reference.cnn which is built from fasta & target
     def (normal_out, normal_args, fasta_args) = ["", "", ""]
-    def target_args = targets ? "--targets $targets" : ""
-    def reference_args = reference ? "--reference $reference" : ""
 
     if (normal_exists){
         def normal_prefix = normal.BaseName
@@ -46,6 +46,9 @@ process CNVKIT_BATCH {
         fasta_args = fasta ? "--fasta $fasta" : ""
     }
 
+    def target_args = targets ? "--targets $targets" : ""
+    def reference_args = reference ? "--reference $reference" : ""
+
     """
     if $is_cram; then
         samtools view -T $fasta $tumor -@ $task.cpus -o $tumor_out

From 1451d62be43f62045d3b30c36a51afacfb3cf1c0 Mon Sep 17 00:00:00 2001
From: SusiJo <susanne.jodoin@gmx.de>
Date: Tue, 24 May 2022 07:38:53 +0200
Subject: [PATCH 12/12] rm md5sums for empty files

---
 tests/modules/cnvkit/batch/test.yml | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/tests/modules/cnvkit/batch/test.yml b/tests/modules/cnvkit/batch/test.yml
index 9bce265a..00d6a767 100755
--- a/tests/modules/cnvkit/batch/test.yml
+++ b/tests/modules/cnvkit/batch/test.yml
@@ -5,7 +5,6 @@
     - cnvkit/batch
   files:
     - path: output/cnvkit/baits.antitarget.bed
-      md5sum: d41d8cd98f00b204e9800998ecf8427e
     - path: output/cnvkit/baits.target.bed
       md5sum: 26d25ff2d6c45b6d92169b3559c6acdb
     - path: output/cnvkit/reference.cnn
@@ -34,7 +33,6 @@
     - cnvkit/batch
   files:
     - path: output/cnvkit/genome.antitarget.bed
-      md5sum: d41d8cd98f00b204e9800998ecf8427e
     - path: output/cnvkit/genome.bed
       md5sum: 87a15eb9c2ff20ccd5cd8735a28708f7
     - path: output/cnvkit/genome.target.bed
@@ -65,7 +63,6 @@
     - cnvkit/batch
   files:
     - path: output/cnvkit/genome.antitarget.bed
-      md5sum: d41d8cd98f00b204e9800998ecf8427e
     - path: output/cnvkit/genome.bed
       md5sum: 87a15eb9c2ff20ccd5cd8735a28708f7
     - path: output/cnvkit/genome.target.bed
@@ -96,7 +93,6 @@
     - cnvkit/batch
   files:
     - path: output/cnvkit/reference.antitarget-tmp.bed
-      md5sum: d41d8cd98f00b204e9800998ecf8427e
     - path: output/cnvkit/reference.target-tmp.bed
       md5sum: 26d25ff2d6c45b6d92169b3559c6acdb
 
@@ -107,6 +103,5 @@
     - cnvkit/batch
   files:
     - path: output/cnvkit/reference.antitarget-tmp.bed
-      md5sum: d41d8cd98f00b204e9800998ecf8427e
     - path: output/cnvkit/reference.target-tmp.bed
       md5sum: 26d25ff2d6c45b6d92169b3559c6acdb