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

Question

[Link]

Hello,

I am trying to run psiblast with a swissprot database in fasta format with

dbname='/path/to/db/swissprot.fasta'
psiblast -comp_based_stats 1 -evalue 0.001 -num_iterations 3 -db $dbname -query /path/to/Q60175.fasta -out_ascii_pssm Q60175.profile

I always get

BLAST Database error: No alias or index file found for protein database [/path/to/db/swissprot.fasta] in search path [/home/main::]

I know this is a possible duplicate to this post but I tried both with swissprot.fasta and only swissprot as well as different folders and path versions and it's always the same error, so I am guessing the error is somewhere else. The swissprot.fasta file is NOT in another folder named swissprot. Does someone know what the '::' at the end of the search path stand means?

Any help is appreciated.

[lieven.sterck]

can you post the output of ls /path/to/db/swissprot* ?

You can also try by setting the BLASTDB env variable on your system to '/path/to/db/' (eg. BLASTDB=/blastdb/shared/ )

[Link]

ls /path/to/db/swissprot* returns /path/to/db/swissprot.fasta

I have tried the shared variation, but unfortunately it results in the same error. I downloaded the fasta file for swissprot from NCBI and the file is written in fasta format, but properties show it as a plain text file, so I renamed it to swissprot.fasta, maybe that is the issue? But I just tried it with renaming it back to only swissprot and that did not work either. Another tool that uses the same fasta file, can use it just fine with the same path.

[lieven.sterck]

well, this likely means you did not format the swissprot DB. Did you run makeblastdb? as an alternative you can also download the pre-formatted version from NCBI

[Link]

Thank you for your help. I have run makeblastdb now, but it results in different files like swissprot.fasta.phr, swissprot.fasta.psi etc. Setting the path to swissprot.fasta.psi for example does not work. I am trying to run the DCMTOP transmembrane protein prediction code and the example path for the DB in the readme is set as following:

1) set the Blast data bank location. e.g. 'set dbname = /home/ThirdPartTools/blast/db/uniref90.fasta'

So I think I have to set my path to the fasta file and not a specific blastdb-file?

[lieven.sterck]

from the look of it it seems your DB is called swissprot.fasta (it's the basename for all the other files makeblastdb has created). So that is the name you have to use in your blast cmdline.

Indeed, any of the files with extensions as .phr. pin ... are files internally used by blast and should thus not be used to call/access the DB

[Link]

Setting it to path/to/db/swissprot,fasta is working now, as long as the blast files are in the same folder, thank you for your help!