Re: Bioinformatics: Regex loop, no output

[tonto]

Maybe like this?

use strict;
use warnings;
use Data::Dumper;

my @proteins=qw(
DAAAAATTLTTTAMTTTTTTCKMMFRPPPPPGGGGGGGGGGGG 
ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD
);
print Dumper \@proteins;
my @new_peptides;

for my $protein (@proteins) {
    if ($protein =~ m/[K(?!P)|R(?!P)]/g) {        
        $protein =~ s/K(?!P)|R(?!P)/=/g;
        push @new_peptides, split ('=',$protein);        
    }
}
print Dumper \@new_peptides;
for (@new_peptides) {
    print "The peptide is $_\n";
}

I use Data::Dumper all the time because I've been fooled by my data too many times. ;-)

[AnomalousMonk]

The output from your code shows some problems:

The peptide is DAAAAATTLTTTAMTTTTTTC
The peptide is MMFRPPPPPGGGGGGGGGGGG
The peptide is ALTAMCMNVWEITYH
The peptide is GSDVN
The peptide is
The peptide is ASFAQPPPQPPPPLLAIKPASDASD

The K or R terminating split codon (if that's the proper term) is being incorrectly removed from the output peptides. (At least, I think this is incorrect. TamaDP doesn't show desired output, but seems satisfied with output examples given in various replies in this thread that include these codons.) So I assume  GSDVN should really be  GSDVNR and the "null" sequence following it should really be the single-codon sequence R. This is all down to the incorrect definition of the  s/// match pattern; take a look at some other replies in this thread for what I feel are more correct  s/// patterns.

In an unrelated note, the regex in the condition expression of the
    if ($protein =~ m/[K(?!P)|R(?!P)]/g) { ... }
block isn't doing what I think you think it's doing. The  [K(?!P)|R(?!P)] character class is exactly equivalent to the  [KPR()?!|] class; metacharacters (alternations, groupings, etc.) have no meaning in a character class, so  ()?!| are just literal characters (and repeated characters have no effect whatsoever). Also, the  /g modifier in the  m//g match is useless in the boolean context of a conditional, although it does no harm (except to burn a few more innocent computrons). Again, all this doesn't affect the basic problem with the code, which stems from the incorrect  s/// match.

I use Data::Dumper all the time because I've been fooled by my data too many times.

Yea and amen brother, yea and amen.

[tonto]

Thank you! I wondered if I understood what was wanted, later posts show that I didn't. I shouldn't have posted that, I'll stop myself next time.