blastn 2.15 support

[muffato]

When switching to blastn 2.15 (tried both 2.15.0--pl5321h6f7f691_0 and _1), there is this warning on stderr:

The -taxids command line option requires additional data files. Please see the section 'Taxonomic filtering for BLAST databases' in https://www.ncbi.nlm.nih.gov/books/NBK569839/ for details.

which suggests we could be doing something wrong, but I can't find on https://www.ncbi.nlm.nih.gov/books/NBK569839/#_usrman_BLAST_feat_Taxonomic_filtering_fo_ any information on making the option work.

In the source code c++/src/algo/blast/blastinput/blast_args.cpp I can find that the warning is printed whenever an exception is raised

        try{ 
            tb.reset(new CTaxonomy4BlastSQLite());
        }    
        catch(CException &){
            LOG_POST(Warning << "The -taxids command line option requires additional data files. Please see the section 'Taxonomic filtering for BLAST databases' in https://www.ncbi.nlm.nih.gov/books/NBK569839/ for details.");
        }    

so it seems there really is something wrong, but I don't know what. It'd be great to get more information about that exception, what it means, where exactly it's raised, etc/.

I also tried downloading a recent NT databases and faced the same problem. Maybe we're missing some files ?

[tkchafin]

Just posting as I think this covers embedding the taxids in custom blastdbs, I can look in more detail when I get a chance later
https://www.ncbi.nlm.nih.gov/books/NBK569841/
https://www.ncbi.nlm.nih.gov/books/NBK569839/#usrman_BLAST_feat.Taxonomic_filtering_fo

[muffato]

We download the NR database from the NCBI and it should already contain the taxon IDs. The problem is that I had the impression that the error was only showing up with 2.15 (same database and same command-line otherwise). Let me find an example

[muffato]

Yes, here is an example

$ cd ~mm49/link/scratch123/blastn/214
$ singularity exec -B /lustre,/nfs /nfs/treeoflife-01/teams/shared/nextflow/cache/nxf_singularity/depot.galaxyproject.org-singularity-blast-2.14.1--pl5321h6f7f691_0.img /bin/bash $PWD/command.sh
Using ./current/nt

$ singularity exec -B /lustre,/nfs /nfs/treeoflife-01/teams/shared/nextflow/cache/nxf_singularity/quay.io-singularity-blast-2.15.0--pl5321h6f7f691_1.img /bin/bash $PWD/command.sh
Using ./current/nt
The -taxids command line option requires additional data files. Please see the section 'Taxonomic filtering for BLAST databases' in https://www.ncbi.nlm.nih.gov/books/NBK569839/ for details.

If I remember correctly, same problem with a fresh download of the NT database at ~mm49/link/scratch123/blastn/db

[tkchafin]

tracking progress in JIRA but adding the solution here for posterity

Apparently we need to stage the tax* files into the working dir. This seems to conflict with what the online docs say, but at least the solution on our end is simple. I'll patch the blastn module to copy these over

(base) tc25@farm22-head2:~/blast$ cp current/tax* .
(base) tc25@farm22-head2:~/blast$ singularity exec -B /lustre,/nfs https://depot.galaxyproject.org/singularity/blast:2.16.0--hc155240_2 /bin/bash $PWD/command.sh
Using ./current/nt

[muffato]

🙌🏼