Tblastx Exception Error

Question

[ES]

Hi,

I am having some problems running tblastx locally on my HPC. I built my own database and I am trying to run tblastx on a small data set (to then move into larger analyses) but I am getting an error (I am using blast 2.2.26)

This is my command to run tblastx:

$ tblastx -num_threads 8 -query nucleotide.fasta -db  nucleotide_db -out tblastx_out.xml -evalue 1e-10 -outfmt 5

And this is the error:

Error: NCBI C++ Exception:
    "/../blast/2.2.26/src/ncbi-blast-2.2.26+-src/c++/src/objects/seq/../seqalign/Seq_align.cpp", line 210: Error: CSeq_align::GetSeqRange(): Row is empty

I used basically the same command with blastp (of course changing the query and database) and it worked fine.

Any help would be much appreciated.

Thanks

[Istvan Albert]

I would double check that query file to make sure it contains what it should

[ES]

Thanks. Yes, it does contain what it should...still getting the same error (on a third attempt)

Answer 1

[bjbliss]

So, your database is a database of ACGT's? Tblastx needs an input database of nucleotides, and it needs a query of nucleotides. blastp uses amino acid database and amino acid query, so I would not expect the same input file to work with both programs.

[ES]

Indeed, the db for tblastx is nucleotides: tblastx -numthreads 8 -query nucleotide.fasta -db nucleotidedb What I meant when I mentioned blastp is that, as I said "of course changing the query and database" it worked fine: blastp -numthreads 8 -query aminoacid.fasta -db aminoaciddb The point I am trying to make is that I used two different blast programs from the same installation of the blast package: blastp worked, tblastx did not.

[Istvan Albert]

I think mentioning blastp kind of threw people off. Since that tool requires both different target and query files it does not provide the information to rule out certain types of problems. I still think it must be some sort of input problem. See if blastn raises an error or not, at least that rules out a few simple problems. Then see if you can translate the nucleotide sequencens for both query and target and whether the content of these conform to what the program expects.

[ES]

I know. I think people were thrown off by including other blast programs. As you suggest, I have run other blast programs (blastp, blastn and blastx) with the appropriate query and target db and all work except tblastx. This suggests to me there is a problem with the translation of the db?