Retrieve the nucleotide sequences initially used from the SeqIDs of the protein sequences (question)

[silvia1234567890]

Retrieve the nucleotide sequences initially used from the SeqIDs of the protein sequences (question)

[shenwei356]

#415 (comment)

[shenwei356]

You can use the option --id-regexp to specify the IDs for matching.

seqkit grep -f nonproductiveIDs.txt nucleotide.fasta -o nonprodnucl.fasta  --id-regexp '^(.+)_frame='

[silvia1234567890]

Hello,
I used that option and I still got this message:

[INFO] 608703 patterns loaded from file

and the file "nonprodnucl.fasta" is empty. Is it okay to use -o for the output files?

Thank you

[shenwei356]

[INFO] 608703 patterns loaded from file

It just shows how many patterns are loaded.

The key problem is that these patterns have suffixes like _frame=1, which do not exist in the sequence file. So, we need to remove these suffixes before searching.

seqkit grep -f <(perl -pne  's/_frame\=\d+$//' nonproductiveIDs.txt) nucleotide.fasta -o nonprodnucl.fasta

Is it okay to use -o for the output files?

It is.

[silvia1234567890]

Oh, I understand.
I would like to create the IDs.txt with the first 15 characters of the headers, because that sequence is an UMI and it's unique for every sequence. How could I do it?

[shenwei356]

seqkit seq nonprod_seq.fasta -ni --id-regexp '(^.+?)\|' -o IDs.txt

Or

seqkit seq nonprod_seq.fasta -ni | cut -d '|' -f 1 > IDs.txt

Then you need to add the same --id-regexp '(^.+?)\|' when using seqkit grep, if the format does not change in nucleotide.fasta.

[silvia1234567890]

and to create the IDs.txt with the full header except the _frame= (example) part? How could I do it?

[shenwei356]

Well, you can just simply combine the methods above.

seqkit seq nonprod_seq.fasta -n | perl -pne  's/_frame\=\d+$//' > IDs.txt

[silvia1234567890]

Hello, I ran this code:
seqkit seq nonproductive_seqs.fasta -ni --id-regexp 's/_frame\=\d+$//' -o IDs.txt
and I'm getting this error:
[ERRO] fastx: regular expression must contain "(" and ")" to capture matched ID. default: ^(\S+)\s?