preformatted bastdb nt volume/alias file missing

Question

[pe_se]

Hi, I just downloaded and decompressed the pre-formatted NCBI nt database, now blastn gives an error message referring to a missing file "nt.58" (which is correct as there are only 57 of them). Can anyone explain whats wrong? Thanks!

update_blastdb.pl --decompress nt

blastn -db ../blastdb/nt -query test.fasta -outfmt 6 -out test.tabular

BLAST Database error: Could not find volume or alias file (nt.58) referenced in alias file (../blastdb/nt). 

Answer 1

[lieven.sterck]

Did you do a md5sum check of the downloaded DB to verify that is is complete (and correct)?

the alias file is a file called DB.nal (or .pal in case of proteins) . That file usually contains just a list of all the parts it needs to pull together to get the complete DB. (should have something like nt.00 , nt.01 ,nt02, ... in it)

This is also linked to the way you format the DB, meaning that theses alias file can be different for the same DB. If I format one and I put the max filesize per chunk lower , then I will end up with more parts and thus my alias file will contain a longer list (bigger filesize, less entries)

[pe_se]

oh ok, thanks a lot! no, I didn't do an md5sum check yet, figuring out how to do that now.. yes, the nt.nal file contains a list up to "nt.58" whereas the files (.ndh, .nhi, etc.) are only up to "nt.57", which seems strange as I downloaded the whole thing at once. Is there a simple way to fix that?

[lieven.sterck]

very simply & hacky is to remove the nt.58 line from the .nal file (though this could result in corrupted DB) .

Indeed, we need to focus on that discrepancy (download & nal file) . Which blast version are you using?

[pe_se]

Its blastn: 2.10.1+ I manually downloaded nt.58.tar.gz right now which seems to have been added right during my download yesterday, however, now I get a different error -
BLAST Database error: Input db vol size does not match lmdb vol size
I am now downloading the whole thing once more and hope for the best...