The BLAST command stops with an error.

Question

[Riku]

Hi all.

I would like to evaluate the assembly quality by running BLAST to the fasta file output by Trinity. I ran the command as follows, but this job was stopped with an error message.

$ ~/miniconda3/envs/py27/bin/blastx \
-query /home/nkarim/avenae/trinity_even_out_dir/Trinity.300.longest.fasta \
-db /home/nkarim/blast/db/nr \
-outfmt 6 \
-evalue 1e-3 \
-out /home/nkarim/blast/output/avenae.out

Error: NCBI C++ Exception:
    T0 "/opt/conda/conda-bld/blast_1626166529662/work/blast/c++/src/corelib/ncbiobj.cpp", line 980: Critical: (CCoreException::eNullPtr) ncbi::CObject::ThrowNullPointerException() - Attempt to access NULL pointer.
     Stack trace:
      /lustre7/home/lustre5/nkarim/miniconda3/envs/py27/bin/../lib/ncbi-blast+/libxncbi.so ???:0 ncbi::CObject::ThrowNullPointerException() offset=0xB7 addr=0x7ffff635daa7
      /lustre7/home/lustre5/nkarim/miniconda3/envs/py27/bin/../lib/ncbi-blast+/libxblast.so ???:0 ncbi::blast::CBlastTracebackSearch::Run() offset=0x9C2 addr=0x7ffff7863ba2
      /lustre7/home/lustre5/nkarim/miniconda3/envs/py27/bin/../lib/ncbi-blast+/libxblast.so ???:0 ncbi::blast::CLocalBlast::Run() offset=0x1659 addr=0x7ffff780acb9
      /home/nkarim/miniconda3/envs/py27/bin/blastx ???:0 CBlastxApp::x_RunMTBySplitDB() offset=0xDEF addr=0x555555576e8f
      /home/nkarim/miniconda3/envs/py27/bin/blastx ???:0 CBlastxApp::Run() offset=0x49 addr=0x555555578689
      /lustre7/home/lustre5/nkarim/miniconda3/envs/py27/bin/../lib/ncbi-blast+/libxncbi.so ???:0 ncbi::CNcbiApplicationAPI::x_TryMain(ncbi::EAppDiagStream, char const*, int*, bool*) offset=0x17B addr=0x7ffff62a3d6b
      /lustre7/home/lustre5/nkarim/miniconda3/envs/py27/bin/../lib/ncbi-blast+/libxncbi.so ???:0 ncbi::CNcbiApplicationAPI::AppMain(int, char const* const*, char const* const*, ncbi::EAppDiagStream, char const*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) offset=0x5FE addr=0x7ffff62a752e
      /home/nkarim/miniconda3/envs/py27/bin/blastx ???:0 main offset=0x77 addr=0x55555556f6a7
      /lib64/libc.so.6 ???:0 __libc_start_main offset=0xF5 addr=0x7ffff5c513d5
      /home/nkarim/miniconda3/envs/py27/bin/blastx ???:0  offset=0x1BFE9 addr=0x55555556ffe9

I'm working on a Linux environment using miniconda. I found some people having similar errors, but they seemed to be having Windows specific errors. What could be wrong with my command?

Thank you very much for your help!

[lieven.sterck]

How much memory did you assign to this job? Newest blast version are quite memory-hungry so it is not unlikely it just ran out of memory.

Can you check the memory (and mem-usage) of that job/process ?

[Riku]

I appreciate for your advices.

I'm using a supercomputer called "DDBJ", and I asigned 64GB memory. Would I be able to run this with more memory?

[lieven.sterck]

yes, in default settings you will need much more mem assigned to that job, try with something like 200Gb and see what that gives

[Michael]

To find out how much memory you have:

cat /proc/meminfo | head

And then get the blastx version:

blastx -version

And check if your blastdb is ok:

blastdbcmd -info /home/nkarim/blast/db/nr

After that, you could try to install an older blast version. Try e.g. 2.9.0 or 2.10.1 Maybe they need a little less memory.

[Riku]

I appreciated your kind help.

I have ran the command according to you.

$ cat /proc/meminfo | head
MemTotal:       48941764 kB
MemFree:         7765348 kB
MemAvailable:   31985704 kB
Buffers:              24 kB
Cached:         24736940 kB
SwapCached:       703004 kB
Active:          6513052 kB
Inactive:       23077048 kB
Active(anon):    2328112 kB
Inactive(anon):  2655972 kB

Although I got this information, this is the information of login node. I was using another node in this job, and I asigned 64GB memory.

$ blastx -version
blastx: 2.12.0+
 Package: blast 2.12.0, build Jul 13 2021 09:03:00

It seemed I have latest version BLAST. I may need to downgrade BLAST with Anaconda.

$ blastdbcmd -info nr
USAGE
  blastdbcmd [-h] [-help] [-db dbname] [-dbtype molecule_type]
    [-entry sequence_identifier] [-entry_batch input_file] [-ipg IPG]
    [-ipg_batch input_file] [-taxids taxonomy_ids] [-taxidlist input_file]
    [-info] [-tax_info] [-range numbers] [-strand strand]
    [-mask_sequence_with mask_algo_id] [-out output_file] [-outfmt format]
    [-target_only] [-get_dups] [-line_length number] [-ctrl_a]
    [-show_blastdb_search_path] [-list directory] [-remove_redundant_dbs]
    [-recursive] [-list_outfmt format] [-exact_length] [-long_seqids]
    [-logfile File_Name] [-version]

DESCRIPTION
   BLAST database client, version 2.12.0+

Use '-help' to print detailed descriptions of command line arguments
========================================================================

Error: Too many positional arguments (1), the offending value: nr
Error:  (CArgException::eSynopsis) Too many positional arguments (1), the offending value: nr

I also used the command you specified, but same error message was outputed. I thought there were too many paths, so I tried this command but an error occurred. In this case, which do you think the cause is BLAST version or database?

[Michael]

Sorry, try

  blastdbcmd -info -db /home/nkarim/blast/db/nr

If that works, then try a downgrade to 2.11. That is working fine for me.

[Riku]

Thank you for your many replies.

It doesn't seem to be working, so it seem that there is the problem in database.

$ blastdbcmd -info -db /home/nkarim/blast/db/nr
BLAST Database error: No alias or index file found for nucleotide database [/home/nkarim/blast/db/nr] in search path [/lustre7/home/lustre5/nkarim/blast::]

I downloaded as follows. Is there are problem? There are the files from "nr.00.tar.gz" to "nr.55.tar.gz" and their decompression files in the downloaded directory.

$ email="my email address" ncbi-blast-dbs nr

[Michael]

I have never used that download script, however make sure the files exist and are where you think they are:

ls -l /home/nkarim/blast/db/nr*

If that shows files, possibly your download is broken anyway, try using update_blastdb.pl for db updates, it might be slower than the other script but is more reliable. Also:

    cd $HOME/blast/db/
    nohup update_blastdb.pl --decompress nr &

and wait.... check nohup.out from time to time. Once you got 56 volume files you can try again.

[Riku]

Your command have worked. Just as you said, the files might be broken.

But I just asigned 600GB memory, and it seem the task is working. If this job goes to the end without a problem, then the problem might had been in the memory.

If an errors occured in this job, I will try to download again according to you. If I still get an error, I hope you can help me again.

Finally, thank you very much for your kind advices everytime!