How to download/extract the same gene from all bacterial genomes on NCBI?

Question

[tbayer]

I would like to get all "groL" gene (nucleotide) sequences from all assembled bacterial genomes on NCBI. For example, searching for (groL) AND "bacteria"[porgn:__txid2] on NCBI Gene get's me a list of the sequences I want, but I seems like I have to open the links one by one to get the seqeuence, the "send to" menu only offers a list. I'd like to avoid sequence searches via BLAST to get this, as they should all be annotated well. Maybe there is a way via the e-utilities?

[Mensur Dlakic]

You may want to try Identical Protein Groups at NCBI:

https://www.ncbi.nlm.nih.gov/ipg/

Type "grol" in the search field:

https://www.ncbi.nlm.nih.gov/ipg/?term=grol

Choose "Send to:" drop-down menu on the right and select Destination -> File, Format -> Fasta.

Sorry, just realized you wanted nucleotides. Maybe this still helps you.

[tbayer]

Unfortunately I need the sequences for a barcoding taxonomy db, so protein (or back-translations) won't help. Thanks!

[GenoMax]

You may have to resort to blasting or perhaps simply back-translating the IPG sequences Mensur Dlakic mentioned below. As you probably saw, in the gene database while there is a list of groL entries not all of them have location information. That means we can't simply extract them based on location.

There may be a datasets way of doing this. Will try to see if I can test.

Answer 1

[tbayer]

So I decided instead of all bacterial genomes it was sufficient to have one per genus, which puts the number at a more manageable ~4k. I used rsync to download the *_cds_from_genomic.fna.gz files for all those from refeseq, and then seqkit grep -n -r -p gene=groL *.gz > groL_only.fasta to extract my gene of interest. The task that remains is to translate the genbank IDs from those sequences to the species, that should work with the e-utilities.