Blastp issue - "No alias or index file found for protein database"

Question

[Corey]

Hi, I am having a strange issue with doing a local BLAST search via bash as part of the Transdecoder-to-Trinotate pipeline. I downloaded the uniref90.fasta, gunzipped, and just ran makeblastdb in my home directory for the sake of simplicity:

makeblastdb -in uniref90.fasta -dbtype prot -parse_seqids -input_type fasta -out uniref

This took maybe 90 minutes and populated the directory with the following files:

• uniref.[00-42].phr
• uniref.[00-42].pin
• uniref.[00-42].pog
• uniref.[00-42].psq
• uniref.pdb
• uniref.pdb-lock
• uniref.ptf
• uniref.ptf-lock

alongside the original .fasta. Next, I try running the following, within that same directory:

blastp -query ~[path to transcriptome transdecoder_dir]/longest_orfs.pep -db uniref -max_target_seqs 1 -outfmt 6 -evalue 1e-5 -num_threads 10 > blastp.outfmt6

And this gives me, invariably, the following error:

BLAST Database error: No alias or index file found for protein database [uniref] in search path [/home_directory::]

I've tried redownloading the fasta, running makeblastdb to output the files to its own folder, running blastp within that folder, running blastp in the home directory and specifying `-db ~/database_folder/uniref', running variations of this same code but specifying file extensions like uniref.pdb or even just uniref90.fasta to see what happens... nothing changes. I installed BLAST in my personal directory on this server about 9 months ago. I ran this pipeline before with no issues. Does anyone have any idea what could be the problem?

[GenoMax]

Can you run blastdbcheck -db uniref -dbtype prot -full and show the result?

[Corey]

blastdbcheck -db uniref -dbtype prot -full
Writing messages to <stdout> at verbosity (Summary)
ISAM testing is ENABLED.
Legacy testing is DISABLED.
TaxID testing is DISABLED.
Using `full' mode: every OID will be tested.

[ERROR] could not find all volume or alias files referenced in uniref, [skipped]
Testing 0 volume(s).
 Result=SUCCESS. No errors reported for 0 volume(s).
Testing 0 alias(es).
 Result=SUCCESS. No errors reported for 0 alias(es).

Total errors: 1

[GenoMax]

Looks like your indexing is either incomplete or you may have moved some of the files around. At this point you will probably want to re-do the indexing and make sure that the job completes without any errors.

[Corey]

so at this point this will be the third time I've indexed from that fasta - in the event I get the same issue, what steps should I try next? should I consider re-downloading uniref90?

[GenoMax]

If you have tried doing this three times then clearly something is going wrong. You need to closely look at the logs of your indexing job to make sure that indexing is working right. Are you assigning enough memory/time for that process? Uniref_90.fa is a 34G compressed file so you may need 60-80G of RAM.

[Corey]

okay, got it working. for some reason the previous makeblastdb runs weren't making all of the index files... I noted when I did this again that those uniref.[00-##] files varied between runs in the ## they went up to. not a clue why, maybe a server-side thing. appreciate all the help

[GenoMax]

Good to hear that you were able to fix the issue.