Blast Can'T Find Databases

Question

[Jesse J]

I've installed BLAST locally, I've downloaded the nr database via update update_blastdb, I've added .ncbirc to the home folder which has

[BLAST]
BLASTDB=/home/bioinformatics/Desktop/blastDB

and yet when I try to run blastn I get this:

blastn -query Desktop/sea_urchin_output1/permutations.fa -db nr
BLAST Database error: No alias or index file found for nucleotide database [nr] in search path [/home/bioinformatics::/home/bioinformatics/Desktop/blastDB:]

I've tried following instructions to get this to run, but it still doesn't work. Is there something I'm missing?

[Neilfws]

What you were told to use and what is correct may not be the same thing :) According to http://www.ncbi.nlm.nih.gov/staff/tao/URLAPI/blastdb.html the non-redundant nucleotide database files are still called nt.*.tar.gz. If you are using blastn then you want nt, not nr (which is proteins).

[Neilfws]

Two thoughts. (1) Have you checked that /home/bioinformatics/Desktop/blastDB contains the required files, their sizes look right and their names all begin with nr.*? (2) Are you sure that nr is the correct database? For blastn, you should be using the nucleotide database, which was named nt the last time I used it.

[Jesse J]

Yes and yes. The files are there and blastn with nr is what I was told to use.

[Michael Schubert]

Just guessing: did you unzip the .tar.gz file?

[arafelyb8]

Nice it worked for me! However I wonder an error message as "blasn expected nucleotides database" to realize i'm not using the correct blast. Thanks

Answer 1

[Torst]

Assuming you have untarred and ungzipped it correctly, the problem is reported correctly, but is not obvious to the beginner:

No alias or index file found for nucleotide database [nr]

You see, "nr" is a PROTEIN database. It's indices are stored in the "nr.pin" file. Note the "p" in pin. Because you are using blastn, which is nucleotide query VS nucleotide database, it is looking for a "nr.nin" index file, not the "nr.pin" you have.

You either need to install "nt", or use "blastx -db nr" instead, which will translate your query in six frames and match against the nr proteins.