/tools/human_genome_variation/linkToGProfile.pl
Perl | 55 lines | 40 code | 7 blank | 8 comment | 4 complexity | ac44394bc1c449d1e413adb287473faa MD5 | raw file
1#!/usr/bin/env perl 2 3use strict; 4use warnings; 5 6################################################### 7# linkToGProfile.pl 8# Generates a link to gprofile for a list of gene IDs. 9# g:Profiler a web-based toolset for functional profiling of gene lists from large-scale experiments (2007) NAR 35 W193-W200 10################################################### 11 12if (!@ARGV or scalar @ARGV != 4) { 13 print "usage: linkToGProfile.pl infile.tab 1basedCol idType outfile\n"; 14 exit 1; 15} 16 17my $in = shift @ARGV; 18my $col = shift @ARGV; 19my $type = shift @ARGV; 20my $out = shift @ARGV; 21 22if ($col < 1) { 23 print "ERROR the column number should be 1 based counting\n"; 24 exit 1; 25} 26my @gene; 27open(FH, $in) or die "Couldn't open $in, $!\n"; 28while (<FH>) { 29 chomp; 30 my @f = split(/\t/); 31 if (scalar @f < $col) { 32 print "ERROR there is no column $col in $in\n"; 33 exit 1; 34 } 35 if ($f[$col-1]) { push(@gene, $f[$col-1]); } 36} 37close FH or die "Couldn't close $in, $!\n"; 38 39my $link = 'http://biit.cs.ut.ee/gprofiler/index.cgi?organism=hsapiens&query=GENELIST&r_chr=1&r_start=start&r_end=end&analytical=1&domain_size_type=annotated&term=&significant=1&sort_by_structure=1&user_thr=1.00&output=png&prefix=TYPE'; 40$link =~ s/TYPE/$type/; 41my $g = join("+", @gene); 42$link =~ s/GENELIST/$g/; 43#print output 44if (length $link > 2048) { 45 print "ERROR too many genes to fit in URL, please select a smaller set\n"; 46 exit; 47} 48open(FH, ">", $out) or die "Couldn't open $out, $!\n"; 49print FH "<html><head><title>g:Profiler link</title></head><body>\n", 50 '<A TARGET=_BLANK HREF="', $link, '">click here to send list of identifiers to g:Profiler</A>', "\n", 51 '</body></html>', "\n"; 52close FH or die "Couldn't close $out, $!\n"; 53 54#also do link that prints text that could be pulled back into Galaxy? 55exit;