This is a test version of Biostars. For the public version, visit https://www.biostars.org.
blast database formatting issue

Hi,

I downloaded the nt nucleotide database using:

$ ./update_blast.pl nt --decompress

Which seemed to work fine - all the files from nt.00 to nt.46 are present. I went to use the database and the same error came up 

 $ blastn -db /home/ucbtsm6/Scratch/database/nt -num_threads 8 -evalue 1e-10 -query /home/ucbtsm6/Scratch/database/deglist.fa -out taxon_blast_ponerine.xml -outfmt 5


BLAST Database error: No alias or index file found for nucleotide database

I then ran 

$ blastdbcheck -db nt -dbtype nucl -verbosity 4

Which gave this output. It seems as though they all PASS on the main body of each subfile (08) then fail when it gets to the metadata for the next subfile (09), but for each one. 

/home/blast/database_index/nt_nucleotide/nt.08 / Sample:       Status for OID 1399110: PASS
  /home/blast/database_index/nt_nucleotide/nt.08 / Sample:       Status for OID 1408039: PASS
  /home/blast/database_index/nt_nucleotide/nt.08 / Sample:       Status for OID 1408850: PASS
 /home/blast/database_index/nt_nucleotide/nt.09
  /home/blast/database_index/nt_nucleotide/nt.09 / MetaData:   [ERROR] caught exception.
  /home/blast/database_index/nt_nucleotide/nt.09 / MetaData: NCBI C++ Exception:
    T0 "/build/ncbi-blast+-bzPf_D/ncbi-blast+-2.2.31/c++/src/objtools/blast/seqdb_reader/seqdbimpl.cpp", line 1477: Error: BLASTDB::ncbi::CSeqDBImpl::GetTaxInfo() - Specified taxid was not found.
Result=FAILURE. 47 errors reported in 47 volume(s).
Testing 1 alias(es).
 /home/blast/database_index/nt_nucleotide/nt.nal
 Result=SUCCESS. No errors reported for 1 alias(es).

Total errors: 47

Anyone know what the 'specified taxid was not found' is all about? I would have thought that using the update_blast.pl script would have automatically included taxids? I've used the automatically downloaded nr protein database before and it worked fine. Perhaps something to do with the decompress option?

Thanks.

blast software error
• 2,984 views • link
updated by Michael Dondrup Michael Dondrup • written by 4galaxy77 4galaxy77

1 answer

You seem to mix up your paths. First, make sure you are using the same directory for all commands:

blastn -db /home/ucbtsm6/Scratch/database/nt [...]

but: 

/home/blast/database_index/nt_nucleotide/nt.08

The second error is likely independent of the first.

Do the following steps, update to the latest version of blast+ (>= v. 2.6)

echo $BLASTDB # if empty, set it
cd $BLASTDB
update_blast.pl nt --decompress
blastdbcheck -db nt -dbtype nucl

And give the output, when that does not work.

Then use blastn like so: blastn -db nt [...]

• 0 views • link
written by Michael Dondrup Michael Dondrup

I was running the first blastn run on a cluster and then checked the database index locally, hence the different paths :)

I didn't realise I was running an old version - the source file says ncbi-blast-2.6.0+ - so I'll update it and try again. Thanks.

• 0 views • link
written by 4galaxy77 4galaxy77

Log in to answer this question.

More posts like this