search

CVUA-RRW / FooDMe

A reproducible and scalable snakemake workflow for the analysis of DNA metabarcoding experiments, with a special focus on food and feed samples.
https://cvua-rrw.github.io/FooDMe
BSD 3-Clause "New" or "Revised" License
4 stars 1 forks source link

[Request] Lock Blast DB Version #68

Open marchoeppner opened 7 months ago

marchoeppner commented 7 months ago

Hi,

I am not 100% sure that I am not misreading this, but in: https://github.com/CVUA-RRW/FooDMe/blob/master/ressources/fetch_nt_blast.sh

it reads as if the script downloads directly out of https://ftp.ncbi.nlm.nih.gov/blast/db/

For the sake of reproducibility, I would suggest to change that to:

https://ftp.ncbi.nlm.nih.gov/blast/db/v5

to make sure that the database version is reproducible - it's always possible that NCBI releases a version 6, which would overwrite the contents in https://ftp.ncbi.nlm.nih.gov/blast/db/.

Cheers, Marc

gregdenay commented 7 months ago

Hi Marc,

you are reading it correctly, this is a welcome enhancement. Thanks a lot for the feedback. It should be noted however that this is the database format Version 5, NOT the content, database content is regulary expanded.

This script could use some improvement to ease the database update process too (it would now redownload the entire database, unless you kept the .tar archives). But I think the BLAST tools could (should) be used directly for this?

Do you want to submit a PR for this ?

Cheers Greg