Filtering fasta file based on identifier

Question

[jyu429]

Hi, I have a fasta file with many segments and I want to filter out all the segments that have a "P" in the identifier of the segment. Is there a conventional way to do so? Thanks.

[jyu429]

Thank you!

Answer 1

[Ram]

bioawk -c fastx '$name ~ /P/ { print ">"$name; print $seq }' <sequences.fa

If you wanna take all except those with a "P",

bioawk -c fastx '$name ! /P/ { print ">"$name; print $seq }' <sequences.fa

bioawk here

[Matt Shirley]

Neat. I've not found a use for bioawk before but this seems perfect.

[Ram]

It kinda clicked out of the blue for me yesterday. Now I'm gonna add this to my arsenal of regular-use tools :-)

[Shahzad]

how can give the transcriptome.fasta and headerlist.txt in this command?

[Ram]

What are those two files?

Answer 2

[Pierre Lindenbaum]

just awk

awk '/^>/{N=0} /^>P/{N=1} {if(N)print}' *.fa

[Rob Syme]

Maybe

/^>\S*P\S*/

To match identifiers (up to the first space) that contain P rather than just identifiers that start with P.

[Ram]

Would this not print only headers, Pierre?

[Pierre Lindenbaum]

no, if there is no 'next' statement, awk continues to scan all the patterns.

[Ram]

Oops, I read it wrong. I read it as the if(N) print being in the same {} as the N=1. My bad!

[jahn.davik]

But where does the ouput go? Sorry for my ignorance.

[Matt Shirley]

"standard out" or "stdout". You can redirect this to a file like:

awk '/^>/{N=0} /^>P/{N=1} {if(N)print}' *.fa > out.fa