bin/bp_seq_length

   1 #!/usr/bin/perl
   2
   3 =head1 NAME
   4
   5 bp_seq_length.pl - lists the number of bases and number of sequences in specified sequence database files
   6
   7 =head1 SYNOPSIS
   8
   9 bp_seq_length.pl *.fa
  10
  11 =head1 DESCRIPTION
  12
  13 bp_seq_length.pl will report the total number of residues and total number of individual sequences contained within a specified sequence database file.
  14
  15 =head1 OPTIONS
  16
  17  -f/--format          - Specify the database format ('fasta' is default).
  18                         This script uses SeqIO and as such formats are
  19                         limited to those which SeqIO system supports.
  20
  21 =head1 FEEDBACK
  22
  23 =head2 Mailing Lists
  24
  25 User feedback is an integral part of the evolution of this and other
  26 Bioperl modules. Send your comments and suggestions preferably to
  27 the Bioperl mailing list.  Your participation is much appreciated.
  28
  29   bioperl-l@bioperl.org                  - General discussion
  30   http://bioperl.org/wiki/Mailing_lists  - About the mailing lists
  31
  32 =head2 Reporting Bugs
  33
  34 Report bugs to the Bioperl bug tracking system to help us keep track
  35 of the bugs and their resolution. Bug reports can be submitted via the
  36 web:
  37
  38   https://github.com/bioperl/bioperl-live/issues
  39
  40 =head1 AUTHOR - Jason Stajich
  41
  42 Jason Stajich E<lt>jason@bioperl.orgE<gt>
  43
  44 =cut
  45
  46 use strict;
  47 use warnings;
  48 use Bio::SeqIO;
  49 use Getopt::Long;
  50
  51 my $format = 'fasta';
  52 GetOptions('f|format:s' => \$format);
  53
  54 exec('perldoc',$0) unless @ARGV;
  55
  56 foreach my $f ( @ARGV ) {
  57     my $in = new Bio::SeqIO(-file => $f,
  58                             -format => $format);
  59     my $len = 0;
  60     my $count = 0;
  61     while( my $seq = $in->next_seq ) {
  62         $len += $seq->length();
  63         $count++;
  64     }
  65
  66     printf "%-10s %d bp $count sequences\n",$f,$len;
  67 }