From ee792b5dbf5217fefb41298df77ce070b2c44782 Mon Sep 17 00:00:00 2001
From: cgpu <chatzipantsiou@gmail.com>
Date: Fri, 10 Jan 2020 03:01:19 +0000
Subject: [PATCH] Adds bwa -k 23 and GenomeChronicler as tool

- [x] Adds -k 23 (bwa mem seed length)
- [x] Exposes as params bwa_cpus, sort_cpus
- [x] Adds GenomeChronicler in tools (sarek logic)
---
 main.nf         | 14 ++++++++------
 nextflow.config |  2 ++
 2 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/main.nf b/main.nf
index 614da20d46..6d656ee32a 100644
--- a/main.nf
+++ b/main.nf
@@ -51,7 +51,7 @@ def helpMessage() {
                                     Available: Mapping, Recalibrate, VariantCalling, Annotate
                                     Default: Mapping
         --tools                     Specify tools to use for variant calling:
-                                    Available: ASCAT, ControlFREEC, FreeBayes, HaplotypeCaller
+                                    Available: ASCAT, ControlFREEC, FreeBayes, HaplotypeCaller, GenomeChronicler
                                     Manta, mpileup, Mutect2, Strelka, TIDDIT
                                     and/or for annotation:
                                     snpEff, VEP, merge
@@ -797,14 +797,14 @@ process MapReads {
     input = hasExtension(inputFile1, "bam") ? "-p /dev/stdin - 2> >(tee ${inputFile1}.bwa.stderr.log >&2)" : "${inputFile1} ${inputFile2}"
     // Pseudo-code: Add soft-coded memory allocation to the two tools, bwa mem | smatools sort
     // Request only one from the user, the other is implicit: 1 - defined
-    bwa_cpus  = Math.floor ( params.bwa_cpus_fraction * task.cpus) as Integer
-    sort_cpus = task.cpus - bwa_cpus
+    bwa_cpus  = params.bwa_cpus  ? params.bwa_cpus  : Math.floor ( params.bwa_cpus_fraction * task.cpus) as Integer
+    sort_cpus = params.sort_cpus ? params.sort_cpus : task.cpus - bwa_cpus
     """
     echo 'bwa_cpus:'  ${bwa_cpus}
     echo 'sort_cpus:' ${sort_cpus}
 
         ${convertToFastq}
-        bwa mem -K 100000000 -R \"${readGroup}\" ${extra} -t ${bwa_cpus} -M ${fasta} \
+        bwa mem -k 23 -K 100000000 -R \"${readGroup}\" ${extra} -t ${bwa_cpus} -M ${fasta} \
         ${input} | \
         samtools sort --threads ${sort_cpus}  - > ${idSample}_${idRun}.bam
     """
@@ -1058,7 +1058,6 @@ process BaseRecalibratorSpark {
 
 
     tag {idPatient + "-" + idSample + "-" + intervalBed.baseName}
-    echo true
 
     input:
         set idPatient, idSample, file(bam), file(bai), file(intervalBed) from bamBaseRecalibrator
@@ -1350,6 +1349,8 @@ process RunGenomeChronicler {
   output:
   file("results_${bam.simpleName}") into chronicler_results
 
+  when: 'genomechronicler' in tools
+
   script:
   
   optional_argument = vep.endsWith("no_vepFile.txt") ? '' : "--vepFile ${vep}"
@@ -3356,7 +3357,8 @@ def defineToolList() {
         'strelka',
         'tiddit',
         'tnscope',
-        'vep'
+        'vep',
+        'genomechronicler'
     ]
 }
 
diff --git a/nextflow.config b/nextflow.config
index 6edcd2e7af..df7843b3fa 100644
--- a/nextflow.config
+++ b/nextflow.config
@@ -77,6 +77,8 @@ params {
   med_resources_fraction = 0.50
   singleCPUMem = 7.GB
   bwa_cpus_fraction   = 0.6
+  bwa_cpus = null
+  sort_cpus = null
 
   // Deprecated params
   annotateVCF = null