A Genbank file is made of a series of Genbank records separated by a line consisting of "//"
You are reading your file line by line: while ($line = <$IN>) and splitting each line in different lines (??) my @line = split(/\///,$line). Obviously, this is not what you want.
An easy way to solve this is to read the file record by record by assigning the value "//" to $/ (see perlvar) and then process the record
Another possibility is to read the file line by line, and update the $accession, $biotype and $sequence variables, accordingly. But you can be in a problem if a record doesn't have one of them
Update: Following the first approach:
use strict;
use warnings;
$/ = "//";
my $genfile = "c:\bemisia_coi.gb";
open my $ifh, "<", $genfile or die "cannot open $genfile: $!\n";
while (my $chunk = <$ifh>){
last if eof $ifh;
my ($accession) = $chunk =~ /LOCUS\s*([A-Z]*\d+)/;
my ($biotype) = $chunk =~ /BIOTYPE\s*([A-Z])/;
my ($sequence) = $chunk =~ /ORIGIN\s*(.*)$/s;
$sequence =~ s/\s|\d//g;
my $outfile = "${accession}_${biotype}";
open my $ofh, '>' $outfile or die "cannot open $outfile: $!\n";
print $ofh, ">$accession\n$sequence";
close $ofh;
}
Additional comments:
- You don't need to lc your input
- I've never seen a "biotype" field in a Genbank file (I can't find it in the specs either), but I may be wrong, so I conserved the code that tries to capture that field, instead you can see the molecule type in the LOCUS field (DNA|PROT)
- A typical "locus name" is made of letters (2 or 3) and numbers, you only test for 8 letters
- The last record of the file is always empty (because the file ends with a "//") line, that is why I use last if eof $ifh;
- Don't forget to close each output file when it is not needed, a Genbank file could be very large and you can run out of open filehandles, they are a finite resource on every operating system
- The "ORIGIN" sub-record is multi-line, so you should add the /s modifier to your regexp (see perlre)
- If you want to scape the "//" in a regular expression, you should escape both slashes: /\/\//
- $sequence =~ s/\s//g; $sequence =~ s/\d//g; can be written $sequence =~ s/\s|\d//g;
