{bio}[foss/2020b] ABRA2 v2.23, CRISPR-DAV v2.3.4, Excel-Writer-XLSX v1.09, FLASH v2.2.00, PRINSEQ v0.20.4, pysamstats v1.1.2 w/ Python 3.8.6

easybuild/easyconfigs/a/ABRA2/ABRA2-2.23-GCC-10.2.0.eb

@@ -0,0 +1,48 @@
+easyblock = 'MakeCp'
+
+name = 'ABRA2'
+version = '2.23'
+
+homepage = 'https://github.com/mozack/abra2'
+description = "Assembly Based ReAligner"
+
+toolchain = {'name': 'GCC', 'version': '10.2.0'}
+
+source_urls = ['https://github.com/mozack/abra2/archive/']
+sources = ['v%(version)s.tar.gz']
+patches = ['ABRA2-2.22_fix-Makefile.patch']
+checksums = [
+    '3993f66a493070ee49df2865b6786a45a0cf6c379bae83e94b8339abbe673289',  # v2.23.tar.gz
+    '05090efb306fc84d09f007a848ce0d0472f8633633b0a6eaf86ab075d092bc0d',  # ABRA2-2.22_fix-Makefile.patch
+]
+
+builddependencies = [('Maven', '3.6.3', '', True)]
+
+dependencies = [
+    ('Java', '11', '', True),
+    ('BWA', '0.7.17'),
+]
+
+parallel = 1
+
+buildopts = 'CXX="$CXX" CXXFLAGS="$CXXFLAGS"'
+buildopts += '&& make standalone CXX="$CXX" CXXFLAGS="$CXXFLAGS"'
+
+files_to_copy = [
+    (['abra'], 'bin'),
+    (['target/libAbra.%s' % SHLIB_EXT], 'lib'),
+    'target/abra2-%(version)s-jar-with-dependencies.jar',
+]
+
+postinstallcmds = ["cd %(installdir)s && mv abra2-%(version)s-jar-with-dependencies.jar abra2-%(version)s.jar"]
+
+sanity_check_paths = {
+    'files': ['abra2-%(version)s.jar', 'bin/abra', 'lib/libAbra.%s' % SHLIB_EXT],
+    'dirs': [],
+}
+
+modextravars = {
+    "LC_ALL": "en_US.UTF-8",
+}
+
+moduleclass = 'bio'

easybuild/easyconfigs/c/CRISPR-DAV/CRISPR-DAV-2.3.4-foss-2020b.eb

@@ -0,0 +1,36 @@
+name = 'CRISPR-DAV'
+version = '2.3.4'
+
+homepage = 'https://github.com/pinetree1/crispr-dav/'
+description = """CRISPR-DAV is a pipeline to analyze
+amplicon-based NGS data of CRISPR clones in a high throughput manner."""
+
+toolchain = {'name': 'foss', 'version': '2020b'}
+
+source_urls = ['https://github.com/pinetree1/crispr-dav/archive/']
+sources = ['v%(version)s.tar.gz']
+patches = ['crispr-dav-2.3.4_remove_hardcoding.patch']
+checksums = [
+    '49975cd48bdbf31fe5a9e2aaa3f5ed85d3cc6f65a422ee3aa8daed890159d2ae',  # v2.3.4.tar.gz
+    # crispr-dav-2.3.4-foss2020b_remove_hardcoding.patch
+    '41475a09754d65cc2c88d4161be6eacdfc98242bccbe49778c4f1f74210a6cb2',
+]
+
+# also needs `naturalsort`, which is now an R extension - see https://github.com/easybuilders/easybuild-easyconfigs/pull/13762
+# if missing, try reinstalling R (https://docs.easybuild.io/en/latest/Partial_installations.html#installing-additional-extensions-using-k-skip)
+dependencies = [
+    ('R', '4.0.3'),
+    ('Perl', '5.32.0'),
+    ('NGS', '2.10.9'),
+    ('Pysam', '0.16.0.1'),
+    ('Excel-Writer-XLSX', '1.09'),
+    ('pysamstats', '1.1.2'),
+    ('ABRA2', '2.23'),
+    ('PRINSEQ', '0.20.4', '-Perl-%(perlver)s'),
+    ('SAMtools', '1.12'),
+    ('BEDTools', '2.30.0'),
+    ('FLASH', '2.2.00'),
+    ('Archive-Zip', '1.68'),
+]
+
+moduleclass = 'bio'

easybuild/easyconfigs/c/CRISPR-DAV/crispr-dav-2.3.4_remove_hardcoding.patch

@@ -0,0 +1,129 @@
+# author: Denis Kristak (INUITS)
+# changing example command to contain absolute paths + tabs escaping + changing plotting type to cairo
+diff -ruN crispr-dav-2.3.4_orig/Examples/example1/run.sh crispr-dav-2.3.4/Examples/example1/run.sh
+--- crispr-dav-2.3.4_orig/Examples/example1/run.sh	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Examples/example1/run.sh	2021-06-22 12:57:41.700453098 +0200
+@@ -4,6 +4,7 @@
+
+ ## If there is problem loading Perl modules., set PERL5LIB accordingly.
+ #export PERL5LIB=$HOME/perlmod/lib/perl5:$PERL5LIB
+-../../crispr.pl --conf conf.txt --region amplicon.bed --crispr site.bed \
+-	--sitemap sample.site --fastqmap fastq.list --genome genomex 
++
++$EBROOTCRISPRMINDAV/crispr.pl --conf $EBROOTCRISPRMINDAV/Examples/example1/conf.txt --region $EBROOTCRISPRMINDAV/Examples/example1/amplicon.bed --crispr $EBROOTCRISPRMINDAV/Examples/example1/site.bed \
++	--sitemap $EBROOTCRISPRMINDAV/Examples/example1/sample.site --fastqmap $EBROOTCRISPRMINDAV/Examples/example1/fastq.list --genome genomex --verbose 1
+
+diff -ruN crispr-dav-2.3.4_orig/Modules/NGS.pm crispr-dav-2.3.4/Modules/NGS.pm
+--- crispr-dav-2.3.4_orig/Modules/NGS.pm	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Modules/NGS.pm	2021-06-21 17:37:39.381993860 +0200
+@@ -296,7 +296,7 @@
+     my $cmd      = "$self->{bwa} mem";
+     $cmd .= " $h{param}" if $h{param};
+     if ( $h{id} && $h{sm} ) {
+-        $cmd .= " -R \'\@RG\tID:$h{id}\tSM:$h{sm}\tPL:$h{pl}\'";
++        $cmd .= " -R \'\@RG\\tID:$h{id}\\tSM:$h{sm}\\tPL:$h{pl}\'";
+     }
+     $cmd .= " $h{idxbase} $h{read1_inf}";
+     $cmd .= " $h{read2_inf}" if -f $h{read2_inf};
+@@ -558,7 +558,7 @@
+     # file is already indexed.
+     my $cmd = "rm -rf $workdir && mkdir -p $workdir" .
+       " && $self->{java} -Djava.io.tmpdir=$tmpdir -jar $h{abra} --threads 2" .
+-      " --ref $h{ref_fasta} --targets $h{target_bed} --working $workdir" .
++      " --ref $h{ref_fasta} --targets $h{target_bed} --tmpdir $workdir" .
+       " --in $h{bam_inf} --out $h{bam_outf}";
+
+     if ($h{single}) {
+diff -ruN crispr-dav-2.3.4_orig/Rscripts/allele.R crispr-dav-2.3.4/Rscripts/allele.R
+--- crispr-dav-2.3.4_orig/Rscripts/allele.R	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Rscripts/allele.R	2021-06-22 13:13:45.868418215 +0200
+@@ -114,7 +114,7 @@
+ } else {
+ 	h<-400
+ 	w<- ifelse(n>40, 13*n, h*1.25)
+-	png(filename=outfile, height=h, width=w)
++	png(filename=outfile, type='cairo', height=h, width=w)
+ }
+
+ on.exit(dev.off())
+diff -ruN crispr-dav-2.3.4_orig/Rscripts/amplicon.R crispr-dav-2.3.4/Rscripts/amplicon.R
+--- crispr-dav-2.3.4_orig/Rscripts/amplicon.R	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Rscripts/amplicon.R	2021-06-22 13:13:45.868418215 +0200
+@@ -111,7 +111,7 @@
+ 	if ( high_res ) {
+ 		tiff(filename=outfile, width=5, height=4, units='in', res=1200)	
+ 	} else {
+-		png(filename=outfile, width=500, height=400)
++		png(filename=outfile, type='cairo', width=500, height=400)
+ 	}
+ 	on.exit(dev.off())
+ 	print(p)
+diff -ruN crispr-dav-2.3.4_orig/Rscripts/hdr.R crispr-dav-2.3.4/Rscripts/hdr.R
+--- crispr-dav-2.3.4_orig/Rscripts/hdr.R	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Rscripts/hdr.R	2021-06-22 13:13:45.868418215 +0200
+@@ -102,7 +102,7 @@
+ 	h<-400
+ 	barspace=60
+ 	w<- ifelse( n*barspace<h, h, n*barspace)
+-	png(filename=outfile, width=w, height=h)
++	png(filename=outfile, type='cairo', width=w, height=h)
+ }
+
+ print(p)
+diff -ruN crispr-dav-2.3.4_orig/Rscripts/indel.R crispr-dav-2.3.4/Rscripts/indel.R
+--- crispr-dav-2.3.4_orig/Rscripts/indel.R	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Rscripts/indel.R	2021-06-22 13:27:15.254201732 +0200
+@@ -81,7 +81,7 @@
+ 	} else {
+ 		h<-450
+ 		w<- ifelse(n>10, 50*n, h)
+-		png(filename=imgfile, height=h, width=w)
++		png(filename=imgfile, type='cairo', height=h, width=w)
+ 	}
+
+ 	on.exit(dev.off())
+diff -ruN crispr-dav-2.3.4_orig/Rscripts/read_chr.R crispr-dav-2.3.4/Rscripts/read_chr.R
+--- crispr-dav-2.3.4_orig/Rscripts/read_chr.R	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Rscripts/read_chr.R	2021-06-22 13:13:45.868418215 +0200
+@@ -72,7 +72,7 @@
+ 	w<-ifelse(n>5, 100*n, 550)
+ 	max_w=1000
+ 	w <- ifelse(w>max_w, max_w, w)
+-	png(filename=outfile, height=h, width=w)
++	png(filename=outfile, type='cairo', height=h, width=w)
+ }
+
+ print(p)
+diff -ruN crispr-dav-2.3.4_orig/Rscripts/read_stats.R crispr-dav-2.3.4/Rscripts/read_stats.R
+--- crispr-dav-2.3.4_orig/Rscripts/read_stats.R	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Rscripts/read_stats.R	2021-06-22 13:13:45.868418215 +0200
+@@ -69,7 +69,7 @@
+ } else {
+ 	h<-400
+ 	w<- ifelse(n>10, 50*n, h)
+-	png(filename=outfile, height=h, width=w)
++	png(filename=outfile, type='cairo', height=h, width=w)
+ }
+
+ p <- ggplot(dat.m, aes(x=Sample, y=value, fill=variable)) + 
+diff -ruN crispr-dav-2.3.4_orig/Rscripts/snp.R crispr-dav-2.3.4/Rscripts/snp.R
+--- crispr-dav-2.3.4_orig/Rscripts/snp.R	2019-07-26 21:15:08.000000000 +0200
++++ crispr-dav-2.3.4/Rscripts/snp.R	2021-06-22 13:13:45.868418215 +0200
+@@ -76,7 +76,7 @@
+ 	if ( high_res ) {
+     	tiff(filename=outfile, width=5, height=4, units='in', res=1200)
+ 	} else {
+-    	png(filename=outfile, width=500, height=400)
++    	png(filename=outfile, type='cairo', width=500, height=400)
+ 	}
+ 	print(p)
+ 	invisible(dev.off())
+@@ -182,7 +182,7 @@
+ 		wt <- 600
+ 	}
+ 	ht <- 400
+-	png(filename=outfile, width=wt, height=ht)
++	png(filename=outfile, type='cairo', width=wt, height=ht)
+ }
+
+ print(p)

easybuild/easyconfigs/e/Excel-Writer-XLSX/Excel-Writer-XLSX-1.09-foss-2020b.eb

@@ -0,0 +1,32 @@
+easyblock = 'PerlModule'
+
+name = 'Excel-Writer-XLSX'
+version = '1.09'
+
+homepage = 'https://metacpan.org/pod/Excel::Writer::XLSX'
+description = """The Excel::Writer::XLSX module can be used to create an Excel file in the 2007+ XLSX format.
+Multiple worksheets can be added to a workbook and formatting can be applied to cells. Text, numbers, and formulas
+can be written to the cells."""
+
+toolchain = {'name': 'foss', 'version': '2020b'}
+
+source_urls = ['https://cpan.metacpan.org/authors/id/J/JM/JMCNAMARA/']
+sources = ['%(name)s-%(version)s.tar.gz']
+checksums = ['d679c6ac19e93c32ab77594c793e41b948c7bb3873b600e70ad637d093dca187']
+
+dependencies = [
+    ('Perl', '5.32.0'),
+    ('Archive-Zip', '1.68'),
+]
+
+options = {'modulename': 'Excel::Writer::XLSX'}
+
+sanity_check_paths = {
+    'files': ['bin/extract_vba'],
+    'dirs': ['lib/perl5/site_perl/%(perlver)s/Excel/Writer/XLSX'],
+}
+
+sanity_check_commands = ['extract_vba --help 2>&1 | grep "This utility is used to extract the VBA project binary from an Excel"']
+
+
+moduleclass = 'tools'

easybuild/easyconfigs/f/FLASH/FLASH-2.2.00-foss-2020b.eb

@@ -0,0 +1,34 @@
+easyblock = 'MakeCp'
+
+name = 'FLASH'
+version = '2.2.00'
+
+homepage = 'https://ccb.jhu.edu/software/FLASH/'
+description = """FLASH (Fast Length Adjustment of SHort reads) is a very fast
+and accurate software tool to merge paired-end reads from next-generation
+sequencing experiments. FLASH is designed to merge pairs of reads when the
+original DNA fragments are shorter than twice the length of reads. The
+resulting longer reads can significantly improve genome assemblies. They can
+also improve transcriptome assembly when FLASH is used to merge RNA-seq data.
+"""
+
+toolchain = {'name': 'foss', 'version': '2020b'}
+
+source_urls = ['https://github.com/dstreett/FLASH2/archive/']
+sources = ['%(version)s.zip']
+checksums = ['1e54b2dd7d21ca3e0595a3ffdd27ef3098f88c4de5b9302ec5ea074b49b79960']
+
+files_to_copy = [(['flash2'], 'bin')]
+
+postinstallcmds = ["cd %(installdir)s/bin && ln -s flash2 flash"]
+
+sanity_check_paths = {
+    'files': ['bin/flash2', 'bin/flash'],
+    'dirs': [],
+}
+
+sanity_check_commands = [
+    "flash --help",
+]
+
+moduleclass = 'bio'

easybuild/easyconfigs/p/PRINSEQ/PRINSEQ-0.20.4-foss-2020b-Perl-5.32.0.eb

@@ -0,0 +1,99 @@
+# author: Denis Kristak
+
+easyblock = 'Tarball'
+
+name = 'PRINSEQ'
+version = '0.20.4'
+versionsuffix = '-Perl-%(perlver)s'
+
+homepage = 'http://prinseq.sourceforge.net'
+description = """A bioinformatics tool to PRe-process and show INformation of SEQuence data."""
+
+toolchain = {'name': 'foss', 'version': '2020b'}
+
+source_urls = ['http://sourceforge.net/projects/prinseq/files/standalone/']
+sources = ['%(namelower)s-lite-%(version)s.tar.gz']
+checksums = ['9b5e0dce3b7f02f09e1cc7e8a2dd77c0b133e5e35529d570ee901f53ebfeb56f']
+
+dependencies = [
+    ('Perl', '5.32.0'),
+    ('cairo', '1.16.0'),
+]
+
+# these are the perl libraries dependencies
+exts_defaultclass = 'PerlModule'
+exts_filter = ("perldoc -lm %(ext_name)s ", "")
+
+exts_list = [
+    ('ExtUtils::Depends', '0.8001', {
+        'source_tmpl': 'ExtUtils-Depends-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/X/XA/XAOC/'],
+        'checksums': ['673c4387e7896c1a216099c1fbb3faaa7763d7f5f95a1a56a60a2a2906c131c5'],
+    }),
+    ('ExtUtils::PkgConfig', '1.16', {
+        'source_tmpl': 'ExtUtils-PkgConfig-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/X/XA/XAOC/'],
+        'checksums': ['bbeaced995d7d8d10cfc51a3a5a66da41ceb2bc04fedcab50e10e6300e801c6e'],
+    }),
+    ('Pod::Usage', '2.0', {
+        'source_tmpl': 'Pod-Usage-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/M/MA/MAREKR/'],
+        'checksums': ['530943a9ac3ba00404d7be8ee8572f30f6db9de123cd725af3647333a87d4fea'],
+    }),
+    ('Cairo', '1.109', {
+        'source_tmpl': 'Cairo-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/X/XA/XAOC/'],
+        'checksums': ['8219736e401c2311da5f515775de43fd87e6384b504da36a192f2b217643077f'],
+    }),
+    ('Statistics::PCA', '0.0.1', {
+        'source_tmpl': 'Statistics-PCA-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/D/DS/DSTH/'],
+        'checksums': ['f8adb10b00232123d103a5b49161ad46370f47fe0f752e5462a4dc15f9d46bc4'],
+    }),
+    ('MIME::Base64', '3.15', {
+        'source_tmpl': 'MIME-Base64-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/G/GA/GAAS/'],
+        'checksums': ['7f863566a6a9cb93eda93beadb77d9aa04b9304d769cea3bb921b9a91b3a1eb9'],
+    }),
+    ('Math::Cephes::Matrix', '0.5305', {
+        'source_tmpl': 'Math-Cephes-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/S/SH/SHLOMIF/'],
+        'checksums': ['561a800a4822e748d2befc366baa4b21e879a40cc00c22293c7b8736caeb83a1'],
+    }),
+    ('Math::MatrixReal', '2.13', {
+        'source_tmpl': 'Math-MatrixReal-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/L/LE/LETO/'],
+        'checksums': ['4f9fa1a46dd34d2225de461d9a4ed86932cdd821c121fa501a15a6d4302fb4b2'],
+    }),
+    ('Text::SimpleTable', '2.07', {
+        'source_tmpl': 'Text-SimpleTable-%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/M/MR/MRAMBERG/'],
+        'checksums': ['256d3f38764e96333158b14ab18257b92f3155c60d658cafb80389f72f4619ed'],
+    }),
+    ('Contextual::Return', '0.2.1', {
+        'source_tmpl': 'Contextual-Return-v%(version)s.tar.gz',
+        'source_urls': ['https://cpan.metacpan.org/authors/id/D/DC/DCONWAY/'],
+        'checksums': ['17a1ed1043a2abe123920894d6023709e834de66ef766be86e4cfbb96b1dd16c'],
+    }),
+]
+
+# this is the script which relies in many extra Perl libraries so we verify it's working fine
+sanity_check_commands = ['%(installdir)s/prinseq-graphs.pl']
+
+modextrapaths = {
+    'PATH': '',
+    'PERL5LIB': 'lib/perl5/site_perl/%(perlver)s/',
+}
+
+postinstallcmds = [
+    "sed -i -e 's|/usr/bin/perl|/usr/bin/env\ perl|' %(installdir)s/*.pl",
+    "chmod +x %(installdir)s/*.pl",
+]
+
+sanity_check_paths = {
+    'files': ['prinseq-lite.pl', 'prinseq-graphs.pl', 'prinseq-graphs-noPCA.pl'],
+    'dirs': []
+}
+
+
+moduleclass = 'bio'