Count of GC in row, but not from N bases

Question

[Korsocius]

Dear all,

I have one problem, it is only one condition for my easy script for counting of GC content per row

awk 'NR>1{n=length($1); gc=gsub("[gcGC]", "", $1); print gc/n}' $i

How to count length of row without N character.

for example:

Input:  ACAGCTTGCNNNN   => length= 9   Gc content=5/9

format of output is not important, only how to count it.

Thanks a lot

[Korsocius]

I think I could do it by

N_count = awk -F N  '{print NF-1}' file

and this result use in

awk 'NR>1{n=length($1); gc=gsub("[gcGC]", "", $1); print gc/(n-$N_count}' $i

Answer 1

[iraun]

This command should work, but it's not in awk, it is in bash.

while read p; do
       len=$(echo $p | sed 's/N//g' | tr -d '\n' | wc -c)
       cnt=$(echo $p | grep -oh 'C\|G\|g\|c' | tr -d '\n' | wc -c)
       gc=$(awk "BEGIN {printf \"%.2f\",${cnt}/${len}}")
       echo -e length:$len --- GC:$gc
done<file

[Korsocius]

Bash is good too, I solve it in awk with bash together. This result is comfortable but only one thing, there is only rounded to hundredths .Thank you..

[iraun]

Glad to help :).

Answer 2

[tomc]

Normalized GC content per row sans N

awk '{gsub("N","");t=length();gsub(/[GC]/,"");print int((t-length())/t*100)/100}'

or, assumes sequence symbols are strictly ACTGN.

awk '{gsub("N","");t=length();gsub(/[AT]/,"");print int(length()/t*100)/100}'