How to rename the fasta header in a multifasta file

Question

[arriyaz.nstu]

Hi, I have a multifasta file like the example below:

>hCoV-19/Bangladesh/BCSIR-NILMRC-523/2021|EPI_ISL_1034736|2021-01-22
CCAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTCTAAACGAACTTTAAAATCTGTGTGG
CGGCTGCATGCTTAGTGCACTCACGCAGTATAATTAATAACTAATTACTGTC
>hCoV-19/Bangladesh/BCSIR-NILMRC-515/2020|EPI_ISL_1034763|2020-12-24
ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTCTAAACG
AATCTGTGTGGCTGTCACTCGGCTGCATGCTTAGTGCACTCACGCAGT
>hCoV-19/Bangladesh/BCSIR-NILMRC-517/2020|EPI_ISL_1035809|2020-12-24
GGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTCTAAACGAACTTTAAAATCTGTGTGGCTG
CTGCATGCTTAGTGCACTCACGCAGTATAATTAATAACTAATTACTGTCGTTGACAGGACACGAGTAACTCG
CTGCAGGCTGCTTACGGTTTCGTCCGTGTTGCAGCCGAT

Here, you can see that each sequence has an id number (as like EPI_ISL_1034736) in the header. I want to keep only the id number in the header. The resulted file will be as like below:

>EPI_ISL_1034736
CCAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTCTAAACGAACTTTAAAATCTGTGTGG
CGGCTGCATGCTTAGTGCACTCACGCAGTATAATTAATAACTAATTACTGTC
>EPI_ISL_1034763
ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTCTAAACG
AATCTGTGTGGCTGTCACTCGGCTGCATGCTTAGTGCACTCACGCAGT
>EPI_ISL_1035809
GGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTCTAAACGAACTTTAAAATCTGTGTGGCTG
CTGCATGCTTAGTGCACTCACGCAGTATAATTAATAACTAATTACTGTCGTTGACAGGACACGAGTAACTCG
CTGCAGGCTGCTTACGGTTTCGTCCGTGTTGCAGCCGAT

Can any of you help me to achieve this? I can use the seqkit replace tool to rename with my own strings. but in this case, I need to keep the sequence id in the header.

[GenoMax]

See if solutions here help: Fasta header trimming

[cpad0112]

$ sed -r '/^>/ s/.*\|(.*)\|.*/>\1/' test.fa

Answer 1

[shenwei356]

You can use a regular expression to capture the ID:

seqkit seq -i --id-regexp  "\|([^\|]+)\|"

Options/Flags used:

     -i, --only-id                   print ID instead of full head

     --id-regexp string              regular expression for parsing ID (default "^(\\S+)\\s?")

Another way: replacing the whole header with captured ID.

seqkit replace -p ".+\|([^\|]+)\|.+"  -r "\$1"