Ygtjki

I have a file with bunch of rows, here is how it looks like (just a head of file):

 "chrom" "startA" "stopA" "genesA" "startB" "stopB" "genesB" "test"
 1 315121 317607 "gene2" 315521 317204 "gene3" 1684
 1 315521 317204 "gene3" 315121 317607 "gene2" 1684
 1 407644 408993 "gene4" 408421 409504 "gene5" 573
 1 407644 408993 "gene4" 408616 410013 "gene6" 378
 1 408421 409504 "gene5" 407644 408993 "gene4" 573
 1 408421 409504 "gene5" 408616 410013 "gene6" 889
 1 408616 410013 "gene6" 407644 408993 "gene4" 378
 1 408616 410013 "gene6" 408421 409504 "gene5" 889
 1 408616 410013 "gene6" 409682 411483 "gene7" 332
....

There are some identical rows (the same pair of genes, just the order of start and stop positions differ, but they are exactly the same) which I need to remove the repeated row.
For example:

1 315121 317607 "gene2" 315521 317204 "gene3" 1684
1 315521 317204 "gene3" 315121 317607 "gene2" 1684

are the same, it is genes 2 and 3 combination just in a different order and I want to remove one of them.

Here is my desired output:

"chrom" "startA" "stopA" "genesA" "startB" "stopB" "genesB" "test"
 1 315121 317607 "gene2" 315521 317204 "gene3" 1684
 1 407644 408993 "gene4" 408421 409504 "gene5" 573
 1 407644 408993 "gene4" 408616 410013 "gene6" 378
 1 408421 409504 "gene5" 408616 410013 "gene6" 889
 1 408616 410013 "gene6" 409682 411483 "gene7" 332

Is there any idea how I can do this task? Thanks

You might try:

awk 'key = $4 < $7 ? $4 SUBSEP $7 : $7 SUBSEP $4 !seen[key]++' file

That stores the minimum necessary to remove the duplicate records.

!seen[key]++ is a "famous" awk idiom to print a record only for the first time "key" is seen.

You can order the triplets of columns 2-3-4 and 5-6-7 by the value in the first column:

perl -lane '@F[1,2,3,4,5,6] = @F[4,5,6,1,2,3] if $F[1] > $F[4]; print "@F"'

Then you can just run sort -u to remove the duplicities (but you need to special case the column names).

Not optimal, but solves the problem:

#!/bin/bash

touch result_genes
dupe_found=0

while read GENE_LINE
do
 GL=$(echo $GENE_LINE | awk 'print $1" "$2" "$3" "$4" "$5" "$6" "$7" "$8')
 while read RESULT_LINE
 do
 RL=$(echo $RESULT_LINE | awk 'print $1" "$5" "$6" "$7" "$2" "$3" "$4" "$8')
 if [ "$GL" == "$RL" ];
 then 
 dupe_found=1
 break
 fi
 done < result_genes

 if [ $dupe_found = 1 ];
 then
 dupe_found=0; 
 else
 echo $GENE_LINE >> result_genes
 fi

done < genes