Bioinformatics: Regex loop, no output

TamaDP has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Bioinformatics: Regex loop, no output by BrowserUk (Patriarch) on Nov 15, 2015 at 17:42 UTC
Neither of the sequences in your array contain a '!' character, so you won't get any output. Perhaps your regex should be `m/(KR)(?!=P)/g` ? With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday' Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". I knew I was on the right track :) In the absence of evidence, opinion is indistinguishable from prejudice.	[reply] [d/l]
Re: Bioinformatics: Regex loop, no output by GrandFather (Saint) on Nov 16, 2015 at 07:04 UTC
The following uses look behind `(?<=...)` with a match set `[KR]`and negative look ahead `(?!P)` that rejects a "following P" match in a split to slice up the protein: `use strict; use warnings; my @proteins = qw( DAAAAATTLTTTAMTTTTTTCKMMFRPPPPPGGGGGGGGGGGG ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD DAAAAATTLTTTAMTTTTTTCK ); for my $protein (@proteins) { my @peptides = split /(?<=[KR])(?!P)/, $protein; next if @peptides < 2; print "Protein: $protein\n"; print "Peptides:\n"; print " $_\n" for @peptides; }` [download] Prints: `Protein: DAAAAATTLTTTAMTTTTTTCKMMFRPPPPPGGGGGGGGGGGG Peptides: DAAAAATTLTTTAMTTTTTTCK MMFRPPPPPGGGGGGGGGGGG Protein: ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD Peptides: ALTAMCMNVWEITYHK GSDVNR R ASFAQPPPQPPPPLLAIKPASDASD` [download] Premature optimization is the root of all job security	[reply] [d/l] [select]
Re^2: Bioinformatics: Regex loop, no output by TamaDP (Initiate) on Nov 16, 2015 at 13:15 UTC
Thanks all! Working now, and I also added more enzymes, selectable by getopts. Peptides printing alright, however, how could I track which peptide comes from which protein in order to make the printing a little more organised? I.e: Protein 1 Peptide 1 Protein 1 Peptide 2 ... Protein n Pepptide x	[reply]
Re^3: Bioinformatics: Regex loop, no output by Not_a_Number (Prior) on Nov 16, 2015 at 15:12 UTC
In what way does GrandFather's solution (which you replied to) not do what you want?	[reply]
Re^4: Bioinformatics: Regex loop, no output by TamaDP (Initiate) on Nov 16, 2015 at 15:37 UTC
Re^5: Bioinformatics: Regex loop, no output by AnomalousMonk (Archbishop) on Nov 16, 2015 at 16:11 UTC
Some notes below your chosen depth have not been shown here
Re: Bioinformatics: Regex loop, no output by tonto (Friar) on Nov 15, 2015 at 18:39 UTC
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 $_\n"; }` [download] I use Data::Dumper all the time because I've been fooled by my data too many times. ;-)	[reply] [d/l]
Re^2: Bioinformatics: Regex loop, no output by AnomalousMonk (Archbishop) on Nov 16, 2015 at 22:32 UTC
The output from your code shows some problems: `The peptide is DAAAAATTLTTTAMTTTTTTC The peptide is MMFRPPPPPGGGGGGGGGGGG The peptide is ALTAMCMNVWEITYH The peptide is GSDVN The peptide is The peptide is ASFAQPPPQPPPPLLAIKPASDASD` [download] The `K` or `R` terminating split codon (if that's the proper term) is being incorrectly removed from the output peptides. (At least, I think this is incorrect. TamaDP doesn't show desired output, but seems satisfied with output examples given in various replies in this thread that include these codons.) So I assume `GSDVN` should really be `GSDVNR` and the "null" sequence following it should really be the single-codon sequence `R`. This is all down to the incorrect definition of the `s///` match pattern; take a look at some other replies in this thread for what I feel are more correct `s///` patterns. In an unrelated note, the regex in the condition expression of the `if ($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. Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]
Re^3: Bioinformatics: Regex loop, no output by tonto (Friar) on Nov 17, 2015 at 21:05 UTC
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.	[reply]
Re: Bioinformatics: Regex loop, no output by james28909 (Deacon) on Nov 15, 2015 at 21:16 UTC
I think this line: `if ($protein =~ m/(KR)!P/g)` [download] is looking for "KR"... NOT "K" or "R". I think what you meant to do was: `if ($protein =~ m/(K\|R)!P/g)` [download]	[reply] [d/l] [select]
Re^2: Bioinformatics: Regex loop, no output by AnomalousMonk (Archbishop) on Nov 16, 2015 at 00:39 UTC
As BrowserUk pointed out here, there is no `!` regex metacharacter/operator (presumably meaning "not-followed-by-the-following-pattern"). What is probably meant is `(?!pattern)` (see Look-Around Assertions in perlre; also see perlretut). This may be what TamaDP was looking for (there are some other good solutions): c:\@Work\Perl\monks>perl -wMstrict -le "my @proteins = qw( DAAAAATTLTTTAMTTTTTTCKMMFRPPPPPGGGGGGGGGGGG ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD AAAKAAA AAAKAAA XXXXXX ); ;; my @new_peptides; for my $protein (@proteins) { if ($protein =~ s{ (?<= [KR]) (?! P) }{=}xmsg) { push @new_peptides, split ('=', $protein); } } ;; for my $peptide (@new_peptides) { print qq{Peptide is '$peptide'}; } " Peptide is 'DAAAAATTLTTTAMTTTTTTCK' Peptide is 'MMFRPPPPPGGGGGGGGGGGG' Peptide is 'ALTAMCMNVWEITYHK' Peptide is 'GSDVNR' Peptide is 'R' Peptide is 'ASFAQPPPQPPPPLLAIKPASDASD' Peptide is 'AAAK' Peptide is 'AAA' Peptide is 'AAAK' Peptide is 'AAA' [download] Un-`s///`-ubstituted input proteins are not `split` and push-ed into the output array. Yet another solution might be: c:\@Work\Perl\monks>perl -wMstrict -le "my @proteins = qw( DAAAAATTLTTTAMTTTTTTCKMMFRPPPPPGGGGGGGGGGGG ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD AAAKAAA AAAKAAA XXXXXX ); ;; my $cleave = qr{ (?<= [KR]) (?! P) }xms; ;; my @peptides = map split($cleave), grep m{ $cleave }xms, @proteins ; ;; print qq{Peptide is '$_'} for @peptides; " Peptide is 'DAAAAATTLTTTAMTTTTTTCK' Peptide is 'MMFRPPPPPGGGGGGGGGGGG' Peptide is 'ALTAMCMNVWEITYHK' Peptide is 'GSDVNR' Peptide is 'R' Peptide is 'ASFAQPPPQPPPPLLAIKPASDASD' Peptide is 'AAAK' Peptide is 'AAA' Peptide is 'AAAK' Peptide is 'AAA' [download] in which the central process might be documented (with the steps of the statement being taken right-to-left, bottom-to-top) as: `my @peptides = # 4. and the pieces are peptides. map split($cleave), # 3. split at each cleavage point... grep m{ $cleave }xms, # 2. that can be cleaved, ... @proteins # 1. for each protein... ;` [download] Update: After sober consideration, I changed the code example above to its present form. The previous code contained `my $cleave = qr{ [KR] (?! P) }xms;` `...` `map split(m{ (?<= $cleave) }xms),` `grep m{ $cleave }xms,` `...` which doesn't really respect the DRY principle, which is what I was aiming to exemplify. Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]
Re: Bioinformatics: Regex loop, no output by Anonymous Monk on Nov 15, 2015 at 17:48 UTC
`#!/usr/bin/perl # http://perlmonks.org/?node_id=1147731 use strict; use warnings; my @new_peptides; my @proteins=qw( DAAAAATTLTTTAMTTTTTTCKMMFRPPPPPGGGGGGGGGGGG ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD ); for my $protein (@proteins) { if ($protein =~ m/[KR](?!P)/) { $protein =~ s/[KR]\K(?!P)/=/g; push @new_peptides, split ('=',$protein); } } for (@new_peptides) { print "The peptide is $_\n"; }` [download]	[reply] [d/l]
Re: Bioinformatics: Regex loop, no output by Laurent_R (Canon) on Nov 15, 2015 at 21:32 UTC
If you don't need to keep the K or R items on which to cut the string, you could try this (demonstrated under the Perl debugger): `DB<6> $p = 'ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD'; DB<7> @b = split /[KR](?!P)/, $p; DB<8> x \@b; 0 ARRAY(0x6004f0f88) 0 'ALTAMCMNVWEITYH' 1 'GSDVN' 2 '' 3 'ASFAQPPPQPPPPLLAIKPASDASD'` [download] If you need to keep those K and R items, then this might be easier: `DB<10> p $p ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD DB<11> $p = 'ALTAMCMNVWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD'; DB<12> $p =~ s/([KR])(?!P)/$1=/g; DB<21> p $p ALTAMCMNVWEITYHK=GSDVNR=R=ASFAQPPPQPPPPLLAIKPASDASD` [download] In the latter case, you can then just split the resulting string on `/=/`.	[reply] [d/l] [select]
Re: Bioinformatics: Regex loop, no output by Anonymous Monk on Nov 15, 2015 at 18:29 UTC
Using the power of map: `#! /usr/bin/perl -wl my @proteins = qw(DAAAAATTLTTTAMTTTTTTCKMMFRPPPPPGGGGGGGGGGGG ALTAMCMN +VWEITYHKGSDVNRRASFAQPPPQPPPPLLAIKPASDASD); my %seen = map {$_ => 1} @proteins; print "Peptide $_" for grep !$seen{$_}++, map {split /[KR]\K(?!P)/} @proteins;` [download]	[reply] [d/l]
Re^2: Bioinformatics: Regex loop, no output by AnomalousMonk (Archbishop) on Nov 15, 2015 at 23:33 UTC
Something I don't understand here. The "uniquifying" action of the `%seen` hash acts to prevent the passage of un-split, whole, original proteins (like `AAAAAA` in the example below) ~~from getting~~ through the dataflow "pipe" into the output, but it also prevents duplicated `split` pieces (e.g., `AAAAK AAAA` below) from the input from being output. Is this bioinformatically useful? `c:\@Work\Perl\monks>perl -wMstrict -MData::Dump -le "my @proteins = qw(AAAAKAAAA AAAAKAAAA AAAAAA); ;; my %seen = map {$_ => 1} @proteins; ;; print qq{Peptide '$_'} for grep !$seen{$_}++, map {split /[KR]\K(?!P)/} @proteins; ;; dd \%seen; " Peptide 'AAAAK' Peptide 'AAAA' { AAAA => 2, AAAAAA => 2, AAAAK => 2, AAAAKAAAA => 1 }` [download] Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]


"be consistent"
	PerlMonks