CODEWITHSUNDEEP
regexstringperl
Robert
Robert

Reputation: 511

Regular expression for a comma separated string

Regular expression for comma separated sub-string permutations

Hi,

I would like to define a regular expression that matches strings consisting of two sub-strings separated by a single comma. Each sub-string may not be empty, and consists exclusively and without repetition of the characters 'A', 'G', 'C' and 'T'. Thus, the pattern should match strings such as:

A,G
AG,CT
TC,CA <- correct, 1st and 2nd sub-strings may have characters in common ( as long as these are not repeated within the sub-string ).
GAT,CGA
CGAT,TG <- correct, sub-strings may be of different length.
etc ...

and should not match:

,G <- missing 1st sub-string
ACGT <- missing comma
X,A <- incorrect character X
AA,G <- repetition of character A in 1st substring
AC,GGC <- repetition of character G in 2nd sub-string
ATGA,TGG <- repetition in both sub-strings
ATCXG,AAC <- incorrect character X and repetition in 2nd sub-string
etc ...

So far I have: 

/^(?=[ACGT]{1,4},[ACGT]{1,4}$)(?!.*(.).*\1.*,)(?!,.*(.).*\1).*$/ 

/^(?=[ACGT]{1,4},[ACGT]{1,4}$)(?!.*(.).*\g{1}.*,)(?!,.*(.).*\g{1}).*$/

also tried joining the capture groups with

/^(?=[ACGT]{1,4},[ACGT]{1,4}$)(?!.*(.).*\g{1}.*,.*(.).*\g{2}).*$/

Now, (?=[ACGT]{1,4},[ACGT]{1,4}$) seems to match the "two sub-strings separated by a single comma" and "consists exclusively of the characters 'A', 'G', 'C' and 'T'" through out the string; (?!.*(.).*\1.*,) seems to match "without repetition" up to the comma.

However, (?!,.*(.).*\1) appears not to be ensuring that it doesn't match a repeated character after the comma.

I'd greatly appreciate replies with clues and/or patterns that help with the desired matching.

Using perl v5.18.2

Thanks in advance

Robert

Upvotes: 1

Views: 728

Answers (2)

user557597
user557597

Reputation:

I think you are pretty close. This should work as well.
It basically does what @Miller's does.

updated - a condensed version. 

 #  /(?m)^(?:(?:^|,)(?:([AGCT])(?![AGCT]*\1)){1,4}){2}$/

 (?m)                   # Multiline mode
 ^                      # BOL
 (?:                    # Total cluster
      (?: ^ | , )            # BOL or comma
      (?:                    # AGCT Cluster grp
           ( [AGCT] )             # (1), Capture single character [AGCT]
           (?!                    # Negative lookahead
                [AGCT]*                # As many [AGCT] needed
                \1                     # to find what is captured in group 1
           )                      # End Negative lookahead
      ){1,4}                 # End  AGCT Cluster grp 1-4 characters
 ){2}                   # Total cluster, do 2 times  
 $                      # EOL

Upvotes: 1

Miller
Miller

Reputation: 35198

Break your problem into steps.

First look for allowed format and characters. Then check to make sure there is no repetition.

use strict;
use warnings;

while (<DATA>) {
    print if /^[ACGT]+,[ACGT]+$/ && !/(\w)\w*\1/;
}

__DATA__
A,G
AG,CT
TC,CA
GAT,CGA
CGAT,TG
,G
ACGT
X,A
AA,G
AC,GGC
ATGA,TGG
ATCXG,AAC

Outputs:

A,G
AG,CT
TC,CA
GAT,CGA
CGAT,TG

Upvotes: 4

Related Questions