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

I am reading the following sequence from a file:

.........RKRMMWW*VWMWRYHDWMH*HR*DRMDMWHMWYVMWVRWMVBHWKVYWSMHYWY*HWVMVS
+KDHMDBYKMWRSMDSD*...**Y*WD*VWDRYHHYRYKRWWDDKDDH*DV**HYW*RW*WMYMRV*BWB

[download]

each character in the sequence stands for something. I am interested in locating '*'. I want to parse out the above sequence (length could be in 100's) into sub-sequences of length in the range 10-25, based on the maximum number of '*' the subsequence could possibly cover, if not then at least should have one '*' in a subsequence string, and finally print the subseq's.

I am just starting in perl..kind of have an idea about the basic search patterns but here I would really appreciate any kind of help!!

  • Comment on How to parse string to substrings based on character occurence in the string
  • Download Code

Replies are listed 'Best First'.
Re: How to parse string to substrings based on character occurence in the string
by ikegami (Patriarch) on Mar 09, 2010 at 05:45 UTC

    Can the subsequences overlap?

    If not, what should 

    aaaaaaaaaaa*aa*aaaaaaaaaaaaaaa*aa*aaaaaaaaaaa

    [download]
    return? The following all match some interpretation of your spec: 
    aaaaaaaaaaa*aa*aaaaaaaa     2
aaaaaaa*aa*aaaaaaaaaaa      2

    [download]
    or 
    aaaaaaaaaaa*aa*aaaaaaaaaa   2
aaaaa*aa*aaaaaaaaaaa        2

    [download]
    or 
    aaaaaaaaaa                  0
a*aa*aaaaaaaaaaaaaaa*aa*a   4
aaaaaaaaaa                  0

    [download]
    or 
    aaaaaaaaaaa                 0
*aa*aaaaaaaaaaaaaaa*aa*     4
aaaaaaaaaaa                 0

    [download]
[reply]
[d/l]
[select]
Re: How to parse string to substrings based on character occurence in the string
by BrowserUk (Patriarch) on Mar 09, 2010 at 05:31 UTC

    What's more important?

    1. That the subsequences contain as many * as possible?
    2. That they are as long as possible? (Or short?)
    3. Must the all subsequences recombine to form the original string?

      Eg. If there is a 50 char section without a *, do you make the subsequences longer to cover it, or omit a bit of the string from the set of substrings?

    4. Is the range 10-25 hard coded, or could you accept 1 or 2 shorter or longer?
    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".
    In the absence of evidence, opinion is indistinguishable from prejudice.
[reply]

      1. yes, the subsequence with as many * possible. that is of main importance

      2. the range is 10-25. as long as the length of the subsequence falls in this range, it is accepted

      3. no, not required.

      If there is a 50 char section without a *, do you make the subsequences longer to cover it--- No.

      Yes- omit a bit of the string from the set of substrings?

      4. yes, only if 1 or 2 shorter or longer.

[reply]
Re: How to parse string to substrings based on character occurence in the string
by BioLion (Curate) on Mar 09, 2010 at 09:47 UTC

    It would really help us to help you in the right direction if you could give us some input and the accompanying expected/desired output. < plug > Also check out Perl and Bioinformatics <\plug>

    Just a something something...
[reply]

      let me clarify my input and output that I want

      input- a sequence 

      .........RKRMMWW*VWMWRYHDWMH*HR*DRMDMWHMWYVMWVRWMVBHWKVYWSMHYWY*HWVMVS
+KDHMDBYKMWRSMDSD*...**Y*WD*VWDRYHHYRYKRWWDDKDDH*DV**HYW*RW*WMYMRV*BWB
+WWDMVSYWDBDWWYSMKW*YRVWVYYRMV*KRK*WWDMRMWR*KR**YWHHWH...DYD*MWKKKKWS

      [download]

      here are the potential subsequences that I am looking for from this sequence:

      Output in the order of the occurrence in the sequence (from left to right)

      
1. *VWMWRYHDWMH*HR* (3 *, 16 length)
2. *HWVMVSKDHMDBYKMWRSMDSD* (2 *, 24 length)
3. *...**Y*WD* (5 *, 11 length)
4. *HDV**HYW*RW*WMYMRV* (6 *, 20 length)
5. *YRVWVYYRMV*KRK*WWDMRMWR* (4*, 25 length)
6. *KRK*WWDMRMWR*KR** (5 *, 18 length)
7. R*KR**YWHHWH...DYD* (4*, 19 length)

      [download]

      does this make sense?

      I do not want the subsequences to overlap necessarily. The importance is more on the number of * than the length. For example, subseq with 6* and 10 length is as good as subseq with 6* and 20 length. So as shorter/longer the subseq could be with max * is accepted (though in the range 10-25)

[reply]
[d/l]
[select]
        does this make sense?

        Not completely.

        1. *HDV**HYW*RW*WMYMRV* doesn't appear in your input.

          Did you mean *DV**HYW*RW*WMYMRV* (6*, 19) or H*DV**HYW*RW*WMYMRV* (6*,20)?

          And if the latter, why?

        2. Why R*KR**YWHHWH...DYD* (4*, 19 length) instead of *KR**YWHHWH...DYD* (4*, 18 length)

        Maybe this is something like your goal? 

        #! perl -slw
use strict;

my $seq = 'RKRMMWW*VWMWRYHDWMH*HR*DRMDMWHMWYVMWVRWMVBHWKVYWSMHYWY*HWVM
+VSKD'
    . 'HMDBYKMWRSMDSD*...**Y*WD*VWDRYHHYRYKRWWDDKDDH*DV**HYW*RW*WMYMRV
+*BWB';

my %uniq;
substr( $seq, $_, 25 ) =~ m[(\*.{8,23}\*)]
    and ++$uniq{ $1 } == 1
    and print "'$1'"
    for 0 .. length( $seq )-1;

__END__
C:\test>827470
'*VWMWRYHDWMH*HR*'
'*HWVMVSKDHMDBYKMWRSMDSD*'
'*...**Y*WD*'
'*WD*VWDRYHHYRYKRWWDDKDDH*'
'*VWDRYHHYRYKRWWDDKDDH*'
'*VWDRYHHYRYKRWWDDKDDH*DV*'
'*DV**HYW*RW*'
'*DV**HYW*RW*WMYMRV*'
'**HYW*RW*WMYMRV*'
'*HYW*RW*WMYMRV*'
'*RW*WMYMRV*'

        [download]
        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".
        In the absence of evidence, opinion is indistinguishable from prejudice.
[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom