0
|
1 #!/usr/bin/perl -w
|
|
2 use strict;
|
|
3 use Bio::SeqIO;
|
|
4
|
|
5 my(@Options, $verbose, $inverse, $file);
|
|
6 setOptions();
|
|
7
|
|
8 my $in = Bio::SeqIO->new(-file=>$file, -format=>'Fasta');
|
|
9 my $out = Bio::SeqIO->new(-fh=>\*STDOUT, -format=>'Fasta');
|
|
10 my $nread=0;
|
|
11 my $nwrote=0;
|
|
12
|
|
13 my $pattern = join('|', @ARGV);
|
|
14
|
|
15 while (my $seq = $in->next_seq) {
|
|
16 $nread++;
|
|
17 my $match = ($seq->description =~ m/($pattern)/ or $seq->display_id =~ m/($pattern)/);
|
|
18 #print STDERR "Found match: ",$seq->display_id, " ", $seq->description, "\n" if $verbose;
|
|
19 if ($match ^ $inverse) { # rare use for XOR !
|
|
20 $out->write_seq($seq);
|
|
21 $nwrote++;
|
|
22 }
|
|
23 }
|
|
24
|
|
25 #print STDERR "Read $nread sequences, wrote $nwrote, with pattern: $pattern\n";
|
|
26 exit(0);
|
|
27 #----------------------------------------------------------------------
|
|
28 # Option setting routines
|
|
29
|
|
30 sub setOptions {
|
|
31 use Getopt::Long;
|
|
32
|
|
33 @Options = (
|
|
34 {OPT=>"h|help", VAR=>\&usage, DESC=>"This help"},
|
|
35 {OPT=>"verbose!", VAR=>\$verbose, DEFAULT=>0, DESC=>"Verbose"},
|
|
36 {OPT=>"v|inverse!", VAR=>\$inverse, DEFAULT=>0, DESC=>"Output NON-matching sequences instead"},
|
|
37 {OPT=>"f|file=s", VAR=>\$file, DEFAULT=>"", DESC=>"The fasta file to extract sequences from"},
|
|
38 );
|
|
39
|
|
40 (!@ARGV) && (usage());
|
|
41
|
|
42 &GetOptions(map {$_->{OPT}, $_->{VAR}} @Options) || usage();
|
|
43
|
|
44 # Now setup default values.
|
|
45 foreach (@Options) {
|
|
46 if (defined($_->{DEFAULT}) && !defined(${$_->{VAR}})) {
|
|
47 ${$_->{VAR}} = $_->{DEFAULT};
|
|
48 }
|
|
49 }
|
|
50 }
|
|
51
|
|
52 sub usage {
|
|
53 print "Usage: $0 [options] id1 [id2 ...] < input.fasta > output.fasta\n";
|
|
54 foreach (@Options) {
|
|
55 printf " --%-13s %s%s.\n",$_->{OPT},$_->{DESC},
|
|
56 defined($_->{DEFAULT}) ? " (default '$_->{DEFAULT}')" : "";
|
|
57 }
|
|
58 exit(1);
|
|
59 }
|
|
60
|
|
61 #----------------------------------------------------------------------
|