Shell Programming and Scripting

Hello,

I have a large file that contains 114 total columns with over 6,000 rows and a header; the final 27 columns are coded in A/T/G/C. There is also a reference column coded A/T/C/G.

e.g. OLD_file

Code:

col1 col2 3 ref ... 27 28 29 30 ...
1 r 22 A ... G A G A ...
2 f 22 C ... T T C T ...
3 g 22 T ... T C T T ...
4 h 22 G ... G G G G ...
.
.
.

I want to create a new file, where the first 26 columns are the same, but I recode the final 27 columns to 0/1 based. Now if the column labeled 'ref' is my reference column, I want to do a logical over the 27 final columns based on the reference column:

For the ith (from column 27 to end) column and jth row (for all rows in the file), if the (i,j)th entry = the ith entry in the reference row of the OLD_file, the (i,j)th entry in the NEW_file = 1 , otherwise the (i,j)th entry in the NEW_file = 0.

So the NEW_file would be recoded:
e.g. NEW_file

Code:

col1 col2 3 ref ... 27 28 29 30 ...
1 r 22 A ... 0 1 0 1 ...
2 f 22 C ... 0 0 1 0 ...
3 g 22 T ... 1 0 1 1 ...
4 h 22 G ... 1 1 1 1 ...
.
.
.

I created a loop in R for this, but it was too slow; I was hoping gawk would be faster-- any ideas?

Thanks a lot in advance!

I don't know if I understood you correctly, but try this:

Code:

awk '{for (i=27;i<=NF;i++)$i=$i==$4?1:0}1' file

It assumes that ref column is the 4th one.

Hey Thanks a lot!