ddocent: Fix trimmomatic path, add gshuf dep

[Jason W Bacon <bacon%NetBSD.org@localhost>]

Module Name:	pkgsrc-wip
Committed By:	Jason W Bacon <bacon%NetBSD.org@localhost>
Pushed By:	outpaddling
Date:		Sun Apr 15 08:06:08 2018 -0500
Changeset:	965fb235cf6aa74b35884aa7638ffd944c940a2d

Modified Files:
	ddocent/Makefile
	ddocent/distinfo
	ddocent/patches/patch-dDocent

Log Message:
ddocent: Fix trimmomatic path, add gshuf dep

To see a diff of this commit:
https://wip.pkgsrc.org/cgi-bin/gitweb.cgi?p=pkgsrc-wip.git;a=commitdiff;h=965fb235cf6aa74b35884aa7638ffd944c940a2d

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.

diffstat:
 ddocent/Makefile              |  1 +
 ddocent/distinfo              |  2 +-
 ddocent/patches/patch-dDocent | 29 +++++++++++++++++++++++++++--
 3 files changed, 29 insertions(+), 3 deletions(-)

diffs:
diff --git a/ddocent/Makefile b/ddocent/Makefile
index 7a6306c239..4da4caa505 100644
--- a/ddocent/Makefile
+++ b/ddocent/Makefile
@@ -22,6 +22,7 @@ LICENSE=	mit
 # ddocent test data do not unpack with FreeBSD 11.1 /usr/bin/unzip
 DEPENDS=	mawk>=0:../../lang/mawk \
 		gawk>=0:../../lang/gawk \
+		coreutils>=0:../../sysutils/coreutils \
 		gnuplot>=0:../../graphics/gnuplot \
 		parallel>=0:../../parallel/parallel \
 		bwa>=0.7.13:../../biology/bwa \
diff --git a/ddocent/distinfo b/ddocent/distinfo
index 34d882ca9d..a2a866e613 100644
--- a/ddocent/distinfo
+++ b/ddocent/distinfo
@@ -4,4 +4,4 @@ SHA1 (dDocent-2.2.25.tar.gz) = 5bdb9419a7d6770114d1bb363fcf1af07d49190b
 RMD160 (dDocent-2.2.25.tar.gz) = a33107f9d3269f647f9970b0bfcd14e15bca9575
 SHA512 (dDocent-2.2.25.tar.gz) = d7ba86d363b50dba648bb7b26e4e3c0306b3b5735e9cdea95a8d00164e12649781f325ae3fe8fba14ed76fa6478ecdf3263b07002cd8df3067fb0359e657765a
 Size (dDocent-2.2.25.tar.gz) = 336804 bytes
-SHA1 (patch-dDocent) = d81a3d26e7ea01305c29d79d4ee688dae9fdef34
+SHA1 (patch-dDocent) = 339a073908c62680c859dc58258ef938469a8d40
diff --git a/ddocent/patches/patch-dDocent b/ddocent/patches/patch-dDocent
index a1e8b393da..5c51384c26 100644
--- a/ddocent/patches/patch-dDocent
+++ b/ddocent/patches/patch-dDocent
@@ -29,8 +29,8 @@ $NetBSD$
  
 -if find ${PATH//:/ } -maxdepth 1 -name TruSeq2-PE.fa 2> /dev/null | grep -q 'Tru' ; then
 -	ADAPTERS=$(find ${PATH//:/ } -maxdepth 1 -name TruSeq2-PE.fa 2> /dev/null | head -1)
-+if [ -e %%PREFIX%%/share/trimmomatic/adapters ]; then
-+       ADAPTERS=%%PREFIX%%/share/trimmomatic/adapters
++if [ -e %%PREFIX%%/share/trimmomatic/adapters/TruSeq2-PE.fa ]; then
++       ADAPTERS=%%PREFIX%%/share/trimmomatic/adapters/TruSeq2-PE.fa
  	else
      echo "The file listing adapters (included with trimmomatic) is not installed or is not in your" '$PATH'"."
      NUMDEP=$((NUMDEP + 1))
@@ -51,6 +51,31 @@ $NetBSD$
  	if [ "$BWAV" -lt "13" ]; then
          	echo "The version of bwa installed in your" '$PATH' "is not optimized for dDocent."
          	echo "Please install at least version 0.7.13"
+@@ -388,10 +392,10 @@ if [ "$SNP" != "no" ]; then
+ 	}
+ 	export -f call_genos
+ 
+-	ls mapped.*.bed | sed 's/mapped.//g' | sed 's/.bed//g' | shuf | parallel --env call_genos --memfree $MAXMemory -j $NUMProc --no-notice call_genos {}
++	ls mapped.*.bed | sed 's/mapped.//g' | sed 's/.bed//g' | gshuf | parallel --env call_genos --memfree $MAXMemory -j $NUMProc --no-notice call_genos {}
+ ####	
+-	#ls mapped.*.bed | sed 's/mapped.//g' | sed 's/.bed//g' | shuf | parallel --memfree $MAXMemory -j $FB1 --no-notice --delay 1 freebayes -L bamlist.list -t mapped.{}.bed -v raw.{}.vcf -f reference.fasta -m 5 -q 5 -E 3 --min-repeat-entropy 1 -V --populations popmap -n 10
+-	#ls mapped.*.bed | sed 's/mapped.//g' | sed 's/.bed//g' | shuf | parallel --memfree $MAXMemory -j $FB1 --no-notice "samtools view -b -L mapped.{}.bed | freebayes -c -t mapped.{}.bed -v raw.{}.vcf -f reference.fasta -m 5 -q 5 -E 3 --min-repeat-entropy 1 -V --populations popmap -n 10"
++	#ls mapped.*.bed | sed 's/mapped.//g' | sed 's/.bed//g' | gshuf | parallel --memfree $MAXMemory -j $FB1 --no-notice --delay 1 freebayes -L bamlist.list -t mapped.{}.bed -v raw.{}.vcf -f reference.fasta -m 5 -q 5 -E 3 --min-repeat-entropy 1 -V --populations popmap -n 10
++	#ls mapped.*.bed | sed 's/mapped.//g' | sed 's/.bed//g' | gshuf | parallel --memfree $MAXMemory -j $FB1 --no-notice "samtools view -b -L mapped.{}.bed | freebayes -c -t mapped.{}.bed -v raw.{}.vcf -f reference.fasta -m 5 -q 5 -E 3 --min-repeat-entropy 1 -V --populations popmap -n 10"
+ 
+ 
+ 	rm mapped.*.bed 
+@@ -447,8 +451,8 @@ fi
+ 
+ #Function for trimming reads using trimmomatic
+ trim_reads(){
+-	TRIMMOMATIC=$(find ${PATH//:/ } -maxdepth 1 -name trimmomatic*jar 2> /dev/null | head -1)
+-    ADAPTERS=$(find ${PATH//:/ } -maxdepth 1 -name TruSeq2-PE.fa 2> /dev/null | head -1)
++       TRIMMOMATIC=%%JAVAJARDIR%%/trimmomatic.jar
++       ADAPTERS=%%PREFIX%%/share/trimmomatic/adapters/TruSeq2-PE.fa
+ 
+ 	if [ -f $1.R.fq.gz ]; then	
+ 		java -Xmx2g -jar $TRIMMOMATIC PE -threads 2 -phred33 $1.F.fq.gz $1.R.fq.gz $1.R1.fq.gz $1.unpairedF.fq.gz $1.R2.fq.gz $1.unpairedR.fq.gz ILLUMINACLIP:$ADAPTERS:2:30:10 LEADING:20 TRAILING:20 SLIDINGWINDOW:5:10 $TW &> $1.trim.log
 @@ -747,7 +751,14 @@ else
  fi