Update: missed the part about multiple I,D sections....so adjusted loop to do that. And now I see that there was some Count Zero code prior to the thread level I've replied to. His code looks fine to me. What I did is very similar except that I used substr() instead of print.

I think this does what you want. Basically in the CIGAR, an insertion becomes a deletion and vice-versa. So I use the edit instructions in the CIGAR in an inverse sense.

The total field lengths in the CIGAR (viewed in inverted sense) may be less than the number of characters in the input, so I think this means truncate the output to whatever that total is.

whether or not some final adjustment to either truncate or perhaps add more "X"'s after inverse of all editing commands is unclear to me - just a matter of knowing what is required - that's why I kept a running tally of the total length.

#!/usr/bin/perl -w
use strict;

while (<DATA>)
{
   next if /^\s*$/;  #skip blank lines
   my ($input, $CIGAR) = split;
   my $ref = $input; #working copy of $input
   
   my (@edit_cmd) = $CIGAR =~ m/\d+\w/g;
   
   my $curr_pos = 0;
   my $total_len =0;
   
   foreach my $cmd (@edit_cmd)
   { 
      if    (my ($M) = $cmd =~ m/(\d+)M/)
      {
         $curr_pos += $M;
         $total_len+= $M;      
      }
   
      elsif (my ($I) = $CIGAR =~ m/(\d+)I/)
      {
         substr($ref,$curr_pos,$I,'');     #delete $I characters
         $total_len -= $I  
      }
   
      elsif (my ($D) = $CIGAR =~ m/(\d+)D/)
      {
         substr($ref,$curr_pos,0,"X" x $D);  #insert $D X's
         $total_len += $D;
         $curr_pos  += $D;
      }  
   }
   
   $ref = substr($ref,0,$total_len); #truncate ?????
   
   print "INPUT = $input   CIGAR = $CIGAR\n";
   print "REF   = $ref\n\n";
   
}

=prints
INPUT = CGAATTAATGGGAATTG   CIGAR = 8M2I7M
REF   = CGAATTAAGGAAT

INPUT = CGAATTAATGGGAATTG   CIGAR = 2M2I2M3D10M
REF   = CGTTTGGGAA

INPUT = CGAATTAATGGGA   CIGAR = 8M2D7M
REF   = CGAATTAAXXTGGGA
=cut

__DATA__
CGAATTAATGGGAATTG 8M2I7M
CGAATTAATGGGAATTG 2M2I2M3D10M
CGAATTAATGGGA     8M2D7M
[download]