Adding columns to a file based on existing columns - perl

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

Answer 2

Here's one way using awk. Run like:
awk -f script.awk file | column -t
Contents of script.awk:
NR==1 {
next
}
length($3)>1 && length($4)==1 {
print $0, $2-1, $2+length($3)+1
next
}
length($3)==1 && length($4)>1 {
print $0, $2, $2+2
next
}1
Results:
chr1 18884 C CAAAA 18884 18886
chr1 135419 TATACA T 135418 135426
chr1 332045 T TTG 332045 332047
chr1 453838 T TAC 453838 453840
chr1 567652 T TG 567652 567654
chr1 602541 TTTA T 602540 602546
chr1 614937 C CTCTCTG 614937 614939
chr1 654889 C CA 654889 654891
chr1 736800 AC A 736799 736803
Alternatively, here's the one-liner:
awk 'NR==1 { next } length($3)>1 && length($4)==1 { print $0, $2-1, $2+length($3)+1; next } length($3)==1 && length($4)>1 { print $0, $2, $2+2; next }1' filem | column -t
The code should be pretty self-explanatory. The 1 on the end of the script simply enables default printing (i.e. '1' returns true) of each line. HTH.

Answer 3

Doing it in perl is trivial (but so is in awk):
#!/usr/bin/perl
while (<>) {
chmop;
my ($chr,$start,$ref,$alt)=split(/\s+/,$_);
if (len($ref) > 1) {
print STDOUT
"$chr\t$start\t$ref\t$alt\t",
$start+len($ref)+1,"\n";
} elsif (len($ref)==1) {
print STDOUT
"$chr\t$start\t$ref\t$alt\t",
$start+2,"\n";
} else {
print STDERR "ERROR: ???\n"; #actually impossible
}
}
Stick it in a file morecols.pl , chmod +x morecols.pl, run more morecols.pl . (Beware, lots of assumptions in this code/instructions). I have a feeling your actual problem is more with programming/text processing then tools or languages. If so, this code is just a stopgap solution....
Cheers.