changeset 1:355e491dbd0f draft

Uploaded
author geert-vandeweyer
date Wed, 08 May 2013 04:34:30 -0400
parents 156964ba18fc
children d03a63a57e82
files VCF_to_VariantDB.pl
diffstat 1 files changed, 134 insertions(+), 0 deletions(-) [+]
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/VCF_to_VariantDB.pl	Wed May 08 04:34:30 2013 -0400
@@ -0,0 +1,134 @@
+#!/usr/bin/perl
+
+# load modules
+use Getopt::Std;
+
+##########################
+# COMMAND LINE ARGUMENTS #
+##########################
+# v = (v)cf file to load
+# V = (V)CF file encoded id
+# u = (u)ser email from galaxy
+# n = sample (n)ame
+# a = sample (a)nnotation
+# g = sample (g)ender
+# o = (o)utput file (simple text file)
+# b = (b)am file (optional)
+# B = (B)am index , needed if b is specified
+# c = encoded id of bam file (optional)
+# C = encoded id of Bam index , needed if b is specified => NOT POSSIBLE YET, NEEDS INDEXING ON VARIANTDB SERVER !
+# S = (S)erver addrress to send data to.
+# R = (r)oot of galaxy web server (/home/galaxyuser/galaxy-dist)
+# H = (H)ost of the galaxy web server (http://my.galaxy.server/galaxy/)
+getopts('v:u:n:a:g:o:b:B:V:c:S:R:H:', \%opts);  # option are in %opts
+
+
+#################
+## CHECK INPUT ##
+#################
+if (!exists($opts{'v'})) {
+	die('No VCF File Specified');
+}
+if (!-e $opts{'v'}) {
+	die('VCF File not found');
+}
+if (!exists($opts{'u'})) {
+	die('No user specified');
+}
+if (!exists($opts{'S'})) {
+	die('No VariantDB server specified');
+}
+if (!exists($opts{'H'})) {
+	die('The Galaxy source-server is not specified');
+}
+
+################
+# open outfile #
+################
+open OUT, ">$opts{'o'}";
+
+###############################
+## TEST CONNECTION TO SERVER ##
+###############################
+use LWP::UserAgent;
+my $url = $opts{'S'}."/";
+$url =~ s/\/\/$/\//;
+$url .= "cgi-bin/galaxy_communication.cgi";
+my $conn = LWP::UserAgent->new();
+my $response = $conn->post( $url, {'HelloWorld' => 1} );
+my $content = $response->decoded_content();
+
+if ($content eq 'HelloGalaxy') {
+	print OUT "Testing connection to $opts{'S'} : OK.\n";
+}
+else {
+	die("Could not connect to the specified server : $content");
+}
+	
+
+##################
+## TEST USER ID ##
+##################
+$email = $opts{'u'};
+my $response = $conn->post( $url, {'CheckUser' => $email} );
+my $content = $response->decoded_content();
+
+if ($content eq 'OK') {
+	print OUT "Testing User-existence : OK.\n";
+}
+else {
+	die("ERROR: $content");
+}
+
+###############################################
+## SEND THE VCF AND BAM FILES FOR PROCESSING ##
+###############################################
+# filepaths
+my $vcfpath = $opts{'v'};
+my $bampath = $opts{'b'};
+my $baipath = $opts{'B'};
+# make output directory in (galaxy/static/) working dir
+my $rand = int(rand(1000));
+our $wd = $opts{'R'}."/static/VCF_parser.".$rand; #int(rand(1000));
+our $dd = $opts{'H'}."/static/VCF_parser.".$rand;
+while (-d $wd) {
+	my $rand = int(rand(1000));
+	$wd = $opts{'R'}."/static/VCF_parser.".$rand;#int(rand(1000));
+	$dd = $opts{'H'}."/static/VCF_parser.".$rand;
+
+}
+$result = system("mkdir $wd");
+
+
+## link files
+$vcfurl = "$dd/data.vcf";
+system ("ln -s $vcfpath $wd/data.vcf");
+if (exists($opts{'b'})) {
+	$bamurl = "$dd/data.bam";
+	$bamidxurl = "$dd/data.bai";
+	system ("ln -s $bampath $wd/data.bam");
+	system ("ln -s $baipath $wd/data.bai");
+}	
+$sample = $opts{'n'};
+$gender = $opts{'g'};
+# post form to the variantDB host. 
+if (exists($opts{'b'})) {
+	$response = $conn->post( $url, {'VCFurl1' => "$vcfurl", 'BAMurl1' => "$bamurl", 'BAIurl1' => "$bamidxurl",'storedata' => 1, 'name1' => "$sample", 'gender1' => "$gender", 'User' => $email} );
+}
+else {
+	$response = $conn->post( $url, {'VCFurl1' => "$vcfurl",  'sample1' => "$sample", 'gender1' => "$gender", 'User' => $email} );
+}
+my $content = $response->decoded_content();
+chomp($content);
+if ($content =~ m/OK$/) {
+	print OUT "Processing Datafiles : OK.\n";
+	print OUT "\n$content\n";
+}
+else {
+	die("ERROR: $content");
+}
+close OUT;
+
+# clean up 
+system("rm -Rf '$wd'");
+