need code for sorting fasta header

Question

[divyaranib.10]

Hello All,

I would like to sort the fasta header line (annotation). Below is the example of how my data is and it is in .txt

>AHF21055.1 ribosomal protein S4 (mitochondrion) [Helianthus annuus]
>AAM96597.1 ATP synthase F0 subunit 6 (mitochondrion) [Chaetosphaeridium globosum]
>AAM96598.1 ATP synthase F0 subunit 8 (mitochondrion) [Chaetosphaeridium globosum]
>AAM96599.1 ATP synthase F0 subunit 9 (mitochondrion) [Chaetosphaeridium globosum]

I would like to get the data as below: just the accession number and protein name preferably in table format and remove everything after the protein name.

example:

>AHF21055.1     ribosomal protein S4
>AAM96597.1     ATP synthase F0 subunit 6
>AAM96598.1     ATP synthase F0 subunit 8
>AAM96599.1     ATP synthase F0 subunit 9

Thank you in advance!

[Eric Lim]

Assuming the (mitochondrion) is always there, this is what I can think on the of my head cut -f1 -d'(' header.txt | sort. There will be an empty space at the end and can be removed by sed 's/ *$//'.

[divyaranib.10]

thank you Eric Lim for your reply!

[Ram]

What have you tried?

PS: Please use the formatting bar (especially the code option) to present your post better. I've done it for you this time.

[divyaranib.10]

Sure Ram will do that from next time. Thanks a lot! I am kinda new to this forum

[Joe]

Do all of your entries follow that format? Will there be some where the string (mitochondrion) is not there?

Answer 1

[n,n]

This gives me the exact output you want as long as (mitochondrion) is present in all lines:

cat old_fasta_headers | sed '/^[[:space:]]*$/d' | cut -d\( -f1 | sed 's/\(\.[[:digit:]]*\) /\1\t/g ; s/$/\n/g' \
> new_fasta_headers

Hope this helps.

[divyaranib.10]

thanks a lot mike!! it solved my problem!

Answer 2

[Ram]

From my experience, FASTA headers consist of two parts - the ID and the description. You can use a tool like bioawk to extract just the identifier and then sort the output, or you can use any combination of command line utilities, such as grep -o or cut or sed, much like Eric Lim's comment.