From a65f5d7aba3a6ebc6f169093cec7fce2c83468b4 Mon Sep 17 00:00:00 2001
From: Brandi Cantarel <brandi.cantarel@utsouthwestern.edu>
Date: Sat, 7 Jul 2018 15:56:22 -0500
Subject: [PATCH] update norm/uniform
---
variants/annotvcf.sh | 0
variants/cnvkit.sh | 0
variants/filter_cnvkit.pl | 0
variants/gatkrunner.sh | 0
variants/germline_vc.sh | 0
variants/norm_annot.sh | 34 ++++++++++++++++++++++++++++++++++
variants/somatic_callers.sh | 0
variants/somatic_vc.sh | 0
variants/svannot.pl | 0
variants/svcalling.sh | 0
variants/uniform_vcf_gt.pl | 16 +++++++++++-----
variants/union.sh | 0
12 files changed, 45 insertions(+), 5 deletions(-)
mode change 100644 => 100755 variants/annotvcf.sh
mode change 100644 => 100755 variants/cnvkit.sh
mode change 100644 => 100755 variants/filter_cnvkit.pl
mode change 100644 => 100755 variants/gatkrunner.sh
mode change 100644 => 100755 variants/germline_vc.sh
create mode 100755 variants/norm_annot.sh
mode change 100644 => 100755 variants/somatic_callers.sh
mode change 100644 => 100755 variants/somatic_vc.sh
mode change 100644 => 100755 variants/svannot.pl
mode change 100644 => 100755 variants/svcalling.sh
mode change 100644 => 100755 variants/union.sh
diff --git a/variants/annotvcf.sh b/variants/annotvcf.sh
old mode 100644
new mode 100755
diff --git a/variants/cnvkit.sh b/variants/cnvkit.sh
old mode 100644
new mode 100755
diff --git a/variants/filter_cnvkit.pl b/variants/filter_cnvkit.pl
old mode 100644
new mode 100755
diff --git a/variants/gatkrunner.sh b/variants/gatkrunner.sh
old mode 100644
new mode 100755
diff --git a/variants/germline_vc.sh b/variants/germline_vc.sh
old mode 100644
new mode 100755
diff --git a/variants/norm_annot.sh b/variants/norm_annot.sh
new file mode 100755
index 0000000..6134c3b
--- /dev/null
+++ b/variants/norm_annot.sh
@@ -0,0 +1,34 @@
+#!/bin/bash
+#union.sh
+
+usage() {
+ echo "-h Help documentation for gatkrunner.sh"
+ echo "-r --Reference Genome: GRCh38 or GRCm38"
+ echo "-p --Prefix for output file name"
+ echo "-v --VCF File"
+ echo "Example: bash union.sh -p prefix -r /path/GRCh38"
+ exit 1
+}
+OPTIND=1 # Reset OPTIND
+while getopts :r:p:v:h opt
+do
+ case $opt in
+ r) index_path=$OPTARG;;
+ p) pair_id=$OPTARG;;
+ v) vcf=$OPTARG;;
+ h) usage;;
+ esac
+done
+function join_by { local IFS="$1"; shift; echo "$*"; }
+shift $(($OPTIND -1))
+baseDir="`dirname \"$0\"`"
+
+source /etc/profile.d/modules.sh
+module load bedtools/2.26.0 samtools/1.6 bcftools/1.6 snpeff/4.3q
+
+prefix="${vcf%.vcf.gz}"
+perl $baseDir\/uniform_vcf_gt.pl $vcf
+bgzip -f ${prefix}.uniform.vcf
+j=${prefix}.uniform.vcf.gz
+tabix -f $j
+bcftools norm -m - -O z -o ${prefix}.norm.vcf.gz $j
diff --git a/variants/somatic_callers.sh b/variants/somatic_callers.sh
old mode 100644
new mode 100755
diff --git a/variants/somatic_vc.sh b/variants/somatic_vc.sh
old mode 100644
new mode 100755
diff --git a/variants/svannot.pl b/variants/svannot.pl
old mode 100644
new mode 100755
diff --git a/variants/svcalling.sh b/variants/svcalling.sh
old mode 100644
new mode 100755
diff --git a/variants/uniform_vcf_gt.pl b/variants/uniform_vcf_gt.pl
index 7dd813e..5f744a4 100755
--- a/variants/uniform_vcf_gt.pl
+++ b/variants/uniform_vcf_gt.pl
@@ -9,6 +9,12 @@ open VCF, "gunzip -c $vcf|" or die $!;
while (my $line = <VCF>) {
chomp($line);
if ($line =~ m/#/) {
+ if ($line =~ m/#CHROM/) {
+ print OUT "##FORMAT=<ID=AO,Number=A,Type=Integer,Description=\"Alternate allele observation count\">\n";
+ print OUT "##FORMAT=<ID=RO,Number=1,Type=Integer,Description=\"Reference allele observation count\">\n";
+ print OUT "##FORMAT=<ID=AD,Number=R,Type=Integer,Description=\"Allelic depths for the ref and alt alleles in the order listed\">\n";
+ print OUT "##FORMAT=<ID=DP,Number=1,Type=Integer,Description=\"Approximate read depth (reads with MQ=255 or with bad mates are filtered)\">\n";
+ }
print OUT $line,"\n";
next;
}
@@ -34,11 +40,6 @@ while (my $line = <VCF>) {
foreach my $i (0..$#deschead) {
$gtdata{$deschead[$i]} = $gtinfo[$i];
}
- if ($gtdata{DP} == 0 || $gtdata{GT} eq './.') {
- push @newgts, '.:.:.:.:.';
- $missingGT ++;
- next FG;
- }
if ($gtdata{AD}){
($gtdata{RO},@alts) = split(/,/,$gtdata{AD});
$gtdata{AO} = join(",",@alts);
@@ -60,6 +61,11 @@ while (my $line = <VCF>) {
if ($gtdata{DP} && $gtdata{DP} < 5) {
$missingGT ++;
}
+ if ($gtdata{DP} == 0 || $gtdata{GT} eq './.') {
+ push @newgts, '.:.:.:.:.';
+ $missingGT ++;
+ next FG;
+ }
push @newgts, join(":",$gtdata{GT},$gtdata{DP},$gtdata{AD},$gtdata{AO},$gtdata{RO});
}
next if ($missingGT == scalar(@gts));
diff --git a/variants/union.sh b/variants/union.sh
old mode 100644
new mode 100755
--
GitLab