Question

I am trying to modify a file which is set up like this:

chr start   ref alt 
chr1    18884   C   CAAAA
chr1    135419  TATACA  T
chr1    332045  T   TTG
chr1    453838  T   TAC
chr1    567652  T   TG
chr1    602541  TTTA    T
chr1    614937  C   CTCTCTG
chr1    654889  C   CA
chr1    736800  AC  A

I want to modify it such that:
if column “ref” is a string >1 (i.e line 2) then I generate 2 new columns where:

first new column = start coordinate-1
second new column = start coordinate+(length of string in ref)+1

therefore, for line 2 output would look like:

chr1 135419 TATACA T 135418 135426

or:
if length of string in “ref” = 1 and column “alt”=string of length>1 (i.e. line 1) then

first new column = start coordinate
second new column = start coordinate+2

so, output for line 1 would be:

chr1 18884 C CAAAA 18884 18886

I have tried to this in awk but without success
My perl is non-existent but would this be the best way? Or maybe in R?

Answer 1

Perl solution. Note that your specification does not mention what to do if both strings are length 1.

#!/usr/bin/perl
use warnings;
use strict;
use feature qw(say);

#use Data::Dumper;
<DATA>; # Skip the header;
while (<DATA>) {
    my ($chr, $start, $ref, $alt) = split;
    my @cols;
    if (1 < length $ref) {
          @cols = ( $start - 1, $start + 1 + length $ref);
    } elsif (1 < length $alt) {
        @cols = ($start, $start + 2);
    } else {
        warn "Don't know what to do at $.\n";
    }
    say join "\t", $chr, $start, $ref, $alt, @cols;
}


__DATA__
chr start   ref alt
chr1    18884   C   CAAAA
chr1    135419  TATACA  T
chr1    332045  T   TTG
chr1    453838  T   TAC
chr1    567652  T   TG
chr1    602541  TTTA    T
chr1    614937  C   CTCTCTG
chr1    654889  C   CA
chr1    736800  AC  A