delete specific columns when other column has specific value (perl or awk) - perl

Question

I have a file with 16 different columns (tab-separated values):
22 51169729 G 39 A 0 0 C 0 0 G 38 0.974359 T 1 0.025641
22 51169730 A 36 A 36 1 C 0 0 G 0 0 T 0 0
22 51169731 C 39 A 0 0 C 39 1 G 0 0 T 0 0
22 51169732 G 37 A 0 0 C 0 0 G 37 1 T 0 0
22 51169733 G 33 A 0 0 C 0 0 G 33 1 T 0 0
22 51169734 C 35 A 0 0 C 35 1 G 0 0 T 0 0
22 51169735 A 32 A 32 1 C 0 0 G 0 0 T 0 0
22 51169736 G 32 A 0 0 C 0 0 G 32 1 T 0 0
22 51169737 C 30 A 0 0 C 30 1 G 0 0 T 0 0
22 51169738 T 27 A 0 0 C 0 0 G 0 0 T 27 1
22 51169739 G 26 A 0 0 C 0 0 G 26 1 T 0 0
22 51169740 A 25 A 25 1 C 0 0 G 0 0 T 0 0
22 51169741 C 22 A 0 0 C 22 1 G 0 0 T 0 0
22 51169742 G 23 A 0 0 C 0 0 G 23 1 T 0 0
22 51169743 C 21 A 0 0 C 21 1 G 0 0 T 0 0
22 51169744 C 22 A 0 0 C 22 1 G 0 0 T 0 0
22 51169745 C 19 A 0 0 C 19 1 G 0 0 T 0 0
22 51169746 C 19 A 0 0 C 19 1 G 0 0 T 0 0
22 51169747 A 15 A 14 0.933333 C 1 0.0666667 G 0 0 T 0 0
22 51169748 C 20 A 0 0 C 20 1 G 0 0 T 0 0
The third column can be A, G, C or T.
I would like to:
remove columns 5, 6 and 7 when column 3 is an 'A' OR when $7=='0'.
Similarly, remove columns 8, 9, 10 when $3== 'C' OR when $10=='0'.
remove columns 11, 12, 13 when $3=='G' OR when $13=='0'.
and remove columns 14, 15, 16 when $3=='T' OR when $16=='0'.
When this is done for the entire file, there would only be 4 columns left in some cases and 7 columns in other cases, like in the following example:
22 51169729 G 39 T 1 0.025641
22 51169730 A 36
22 51169731 C 39
22 51169732 G 37
22 51169733 G 33
22 51169734 C 35
22 51169735 A 32
22 51169736 G 32
22 51169737 C 30
22 51169738 T 27
22 51169739 G 26
22 51169740 A 25
22 51169741 C 22
22 51169742 G 23
22 51169743 C 21
22 51169744 C 22
22 51169745 C 19
22 51169746 C 19
22 51169747 A 15 C 2 0.133333
22 51169748 C 20
Any suggestions?

Answer 1

Perl solution for the first part:
#!/usr/bin/perl
use warnings;
use strict;
my %remove = ( A => 4, # Where to start removing the columns
C => 7, # for a given character in column #3.
G => 10,
T => 13,
);
$\ = "\n"; # Add newline to prints.
$, = "\t"; # Separate values by tabs.
while (<>) { # Read input line by line;
chomp; # Remove newline.
my #F = split /\t/; # Split on tabs, populate an array.
splice #F, $remove{ $F[2] }, 3; # Remove the columns.
print #F; # Output.
}
Once you clarify the second requirement, I can try to add more code. What values do you want to remove? Can you show more examples?

Answer 2

Here's one way to do the first part, assuming no empty fields:
$ cat tst.awk
$3 == "A" { $5=$6=$7="" }
$3 == "C" { $8=$9=$10="" }
$3 == "G" { $11=$12=$13="" }
$3 == "T" { $14=$15=$16="" }
{ gsub(/[[:space:]]+/,"\t"); print }
$ awk -f tst.awk file
1 957584 C 157 A 1 0.006 G 0 0 T 0 0
I don't really understand what you're trying to do in the 2nd part but it sounds like this might be what you want if the test on $7/10/13 is the modified field numbers after the first phase:
$3 == "A" { $5=$6=$7="" }
$3 == "C" { $8=$9=$10="" }
$3 == "G" { $11=$12=$13="" }
$3 == "T" { $14=$15=$16="" }
{ $0=$0 }
$7 ~ /0/ { c++ }
$10 ~ /0/ { c++ }
$13 ~ /0/ { c++ }
c > 1 { $8=$9=$10="" }
{ c=0; gsub(/[[:space:]]+/,"\t"); print }
or this if the test on $7/10/13 is the original field numbers:
$7 ~ /0/ { c++ }
$10 ~ /0/ { c++ }
$13 ~ /0/ { c++ }
$3 == "A" { $5=$6=$7="" }
$3 == "C" { $8=$9=$10="" }
$3 == "G" { $11=$12=$13="" }
$3 == "T" { $14=$15=$16="" }
c > 1 { $8=$9=$10="" }
{ c=0; gsub(/[[:space:]]+/,"\t"); print }
If not, edit your question to clarify with a better example.