Ttyjfyk

I have a file that looks like this

 ID A1 A2 A3
 1 A G A
 2 T G A
 3 T A G
 4 T G A
 5 A A G
 6 A C A
 7 C T G

which is thousands of rows long and made up of G,C,T,A, where G complements C and A complements T. What I'm trying to do is to search for a match for A1 in either in A2 or A3. If there is a match then for it to be left as it is and if there isn't to change A2 and A3 to their complements i.e. A=T and G=C and vice versa.

So the output would be:

 ID A1 A2 A3
 1 A G A
 2 T C T
 3 T T C
 4 T C T
 5 A A G
 6 A C A
 7 C A C

I thought I could do it by using awk to filter matching and unmatching IDs using these:

 awk '' mergedlist > nonmatchlist

and

 awk '' mergedlist > matchlist

but it only worked for one variable i.e T in the former and A in the latter.

perl -lane 'sub flip if ($_[0] eq "T") "A" elsif ($_[0] eq "A") "T" elsif ($_[0] eq "G") "C" elsif ($_[0] eq "C") "G" else $_[0] if (!($F[1] eq $F[2] or $F[1] eq $F[3])) $F[2] = flip($F[2]); $F[3] = flip($F[3]) print "@F"' < input

Should be easy to port back to awk as it's not really doing anything fancy, but that would take me more time to figure out.

You could construct an associative array as a lookup table for the complements e.g.

awk '
 BEGIN 
 complement["A"]="T"; complement["T"]="A";
 complement["C"]="G"; complement["G"]="C";
 

 NR>1 && $3!=$2 && $4!=$2 
 $3 = complement[$3]; 
 $4 = complement[$4];
 

 
 print;
 
' file

Alternatively to an array as suggested by @steeldriver, you could define a function:

awk '
 BEGIN FS == " +" 
 NR == 1 print $0 
 function CHANGE( F )
 
 if ( F == "A" ) F = "T"
 else if ( F == "T" ) F = "A"
 else if ( F == "C" ) F = "G"
 else F = "C"
 return F
 
 NR >= 2 
' file