Mercurial > repos > yufei-luo > s_mart

comparison SMART/bacteriaRegulatoryRegion_Detection/coverageGff.pl @ 18:94ab73e8a190

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author m-zytnicki
date Mon, 29 Apr 2013 03:20:15 -0400
parents
children
comparison
equal deleted inserted replaced
17:b0e8584489e6 18:94ab73e8a190
1 #!/usr/bin/perl -w
2 ###
3 # But : ajout ou modif de la couverture d'un gff
4 #
5 # Entrees : fichier gff
6 #
7 # Sortie : gff affiche a l'ecran
8 #
9 ###------------------------------------------------------
10
11 #!/usr/bin/perl -w
12
13 use vars qw($USAGE);
14 use strict;
15
16 =head1 NAME
17
18 coverageGff.pl - add or compute the coverage of a gff file
19
20 =head1 SYNOPSIS
21
22 % coverageGff.pl -i file.gff -l readLength [-h]
23
24 =head1 DESCRIPTION
25 This script will parse gff file, compute read coverage form the "nbElements" tag and write coverage in gff3 format.
26
27 -i|--input fileName gff input file name
28 -l|--length ReadLength lenght of the reads in bp [38 default]
29 -o|--output fileName gff3 output file name
30 [-h|--help] help mode then die
31
32 =head1 AUTHOR - Claire Toffano-Nioche - fev.11
33
34 =cut
35 #-----------------------
36 my ($fileName, $length, $outFileName) = ("", 38, "coverageOut.gff3") ;
37 # command line check
38 foreach my $num (0 .. $#ARGV) {
39 SWITCH: for ($ARGV[$num]) {
40 /--input|-i/ && do {
41 $fileName=$ARGV[$num+1];
42 open ( fichierGff, "< $fileName" ) or die "Can't open gff file: \"$fileName\"\n" ;
43 last };
44 /--length|-l/ && do {
45 $length=$ARGV[$num+1];
46 last };
47 /--help|-h/ && do { exec("pod2text $0\n") ; die };
48 }
49 }
50 # informations retrieval
51 # open(OUT,">$outFileName") or die "Error can't $outFileName open for output. $!\n";
52 my @lines = <fichierGff> ;
53 close fichierGff ;
54 # treatment
55 #print "gff file read ; number of lines : $#lines\n";
56 for (my $i=0 ; $i <= $#lines ; $i++) {
57 # compute coverage :
58 if ($lines[$i] =~ /nbElements=/) {
59 my ($nbE)=($lines[$i] =~ /nbElements=(\d+)/) ;
60 my @gffCol=split("\t", $lines[$i]) ;
61 # print "ligne : $i, nbE : $nbE, length : $length, debut : $gffCol[3], fin : $gffCol[4].\n";
62 my $cover=$length*$nbE/($gffCol[4]-$gffCol[3]+1) ;
63 $cover=int(100*$cover+0.5)/100 ; # arronri sup. precision 2 chiffres
64 if ($lines[$i] =~ /coverage=/) { # replace coverage
65 $lines[$i] =~ s/coverage=.*;/coverage=$cover;/ ;
66 } else { # add coverage
67 $lines[$i] =~ s/;/;coverage=$cover;/ ;
68 }
69 }
70 # print OUT $lines[$i] ;
71 print $lines[$i] ;
72 }
73 #close OUT ;
74 exit(0);