igBLAST query/options error

Question

[moran.an]

When I try to run this command:

igblastn -germline_db_V $GERMLINE_DB"/human_gl_HV" \
    -germline_db_J $GERMLINE_DB"/human_gl_HJ" \
    -germline_db_D $GERMLINE_DB"/human_gl_HD" \
    -organism human -domain_system imgt \
    -query $WORKDIR"/"$FILE".fasta" \
    -auxiliary_data $IGBLASTDIR"/optional_file/human_gl.aux" \
    -outfmt 7 -num_threads 4 \
    -num_alignments_V 5 \
    -out $FILE"_tab.igblast"

I get this error:

BLAST query/options error: Germline annotation database human/human_V could not be found in [internal_data] directory

(*assume that all the variables are found ok.)

There is a folder named internal_data inside the directory /home/user/igblast/ and there is an imgt germline files

What should I do to fix it?

Thanks

[Ram]

You might wanna substitute appropriate values for the variables - that would make the problem clearer for all of us. Also, you might wanna use this syntax to demarkate variable names from surrounding literal strings:

"${variable_name}LITERAL{another_variable}"

That would make your command a bit clearer as well:

igblastn -germline_db_V "${GERMLINE_DB}/human_gl_HV" \
    -germline_db_J "${GERMLINE_DB}/human_gl_HJ" \
    -germline_db_D "${GERMLINE_DB}/human_gl_HD" \
    -organism human -domain_system imgt \
    -query "{$WORKDIR}/{$FILE}.fasta" \
    -auxiliary_data "{$IGBLASTDIR}/optional_file/human_gl.aux" \
    -outfmt 7 -num_threads 4 \
    -num_alignments_V 5 \
    -out "{$FILE}_tab.igblast"

[moran.an]

Thank you. I will try to make it clearer:

#directory of igblastn
IGBLASTDIR='/home/userlab/igblast'
#directory of IMGT files
GERMLINE_DB='home/userlab/igblast/IMGT_db_for_Igblast'

/home/userlab/igblast/igblastn \
    -germline_db_V "${GERMLINE_DB}/human_gl_HV" \
    -germline_db_J "${GERMLINE_DB}/human_gl_HJ" \
    -germline_db_D "${GERMLINE_DB}/human_gl_HD" \
    -organism human -domain_system imgt \
    -query "${WORKDIR}/${FILE}.fasta" \
    -auxiliary_data "${IGBLASTDIR}/optional_file/human_gl.aux" \
    -outfmt 7 -num_threads 4 \
    -num_alignments_V 5 \
    -out "${FILE}_tab.igblast"

Answer 1

[camole]

I have struggled with the issue and finally able to solve it. There are several aspects to it:

1. Make sure you have the correct directory structure.
2. Need to export IGDATA=path/to/bin/folder. Note that the IGDATA should point to the bin folder and not the internal_data folder. If you're running conda (or an virtual environment), do NOT add IGDATA as the virtual environment variable, like what you do with conda env config vars. Directly export it.
3. For IGDATA, supply an absolute path (from the root) instead of a relative path (from a certain folder)
4. Your absolute path should not contain spaces (might as well avoid _ and - in the directory names as well).

For detail on database setup, this might be helpful: https://github.com/xinyu-dev/igblast

[tennysonsrock.97]

I would like to add to @camole and say that you should move the internal_data directory to the bin folder, where the IgBLAST executables (igblastn, igblastp, etc.) reside.

That solved the problem for me!

[Ram]

Moving a data directory to a binaries directory is not good practice at all. it might work as a hack, but please don't suggest as a solution.

Answer 2

[mikhail.shugay]

Hello!

Are you using the latest IgBlast version from here: ftp://ftp.ncbi.nih.gov/blast/executables/igblast/release/1.4.0/ ? It is hard to know exactly what is the problem without full directory structure and lists of files that are there. I could only suggest setting the environmental variable $IGDATA to point to the dir where the internal_data/ resides.

I've been struggling a while with making IgBlast works, so I've created a completely out-of-box IgBlastWrapper application, which could be of use for you. Its main focus are antibody sequences which cover CDR3 thus allowing clonotype assembly. If you don't have CDR3 covered, you can either skip -c option to get V segment table, or use a --debug mode to get plain IgBlast output, i.e.

java -Xmx40G -jar igblastwrp.jar --debug -R IGH --all-alleles input.fasta - > output.txt

Hope this helps!

[patel.sonal80]

Hi Mikhail. Can IgBlastWrapper be used for species other than those specified in the help page (human, mouse, rat, rabbit, rhesus_monkey)? I'd like to use some pig Ig sequences as the database. Thanks