Entering edit mode
6.7 years ago
1291016966
•
0
Hi all, Recently, I run blastn on ubuntu. The command just as below:
nohup /analysis/Software/ncbi-blast-2.6.0+/bin/blastn -query data/S08-C-1-K.1.fasta -db /database/blast_db/h_abfpv_genome -out blastn_m6.out -outfmt 6 -evalue 1e-5 -num_threads 10 -max_target_seqs 1 &
But there is an error.
Error: NCBI C++ Exception:
T0 "/home/coremake/release_build/build/PrepareRelease_Linux64-Centos_JSID_01_350334_130.14.22.10_9008__PrepareRelease_Linux64-Centos_1481139955/c++/compilers/unix/../../src/corelib/ncbiobj.cpp", line 977: Critical: ncbi::CObject::ThrowNullPointerException() - Attempt to access NULL pointer.
T0 "/home/coremake/release_build/build/PrepareRelease_Linux64-Centos_JSID_01_350334_130.14.22.10_9008__PrepareRelease_Linux64-Centos_1481139955/c++/compilers/unix/../../src/app/blast/blast_app_util.cpp", line 766: Critical: BLAST::ncbi::BlastFormatter_PreFetchSequenceData() - Error pre-fetching sequence data
Stack trace:
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::CStackTrace::CStackTrace(std::string const&) offset=0x8B addr=0x1a9587b
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::CException::SetSeverity(ncbi::EDiagSev) offset=0xEB addr=0x1bc0d8b
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::CObject::ThrowNullPointerException() offset=0x2EA addr=0x1c2db5a
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::objects::CBioseq_Handle::IsSetInst_Fuzz() const offset=0x0 addr=0x143e9a0
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::blast::LoadSequencesToScope(std::vector<ncbi::objects::CSeq_id_Handle, std::allocator<ncbi::objects::CSeq_id_Handle> >&, std::vector<ncbi::CRange<unsigned int>, std::allocator<ncbi::CRange<unsigned int> > >&, ncbi::CRef<ncbi::objects::CScope, ncbi::CObjectCounterLocker>&) offset=0x3BF addr=0xcc93ff
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::BlastFormatter_PreFetchSequenceData(ncbi::blast::CSearchResultSet const&, ncbi::CRef<ncbi::objects::CScope, ncbi::CObjectCounterLocker>, ncbi::blast::CFormattingArgs::EOutputFormat) offset=0x82 addr=0x990d52
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 CBlastnApp::Run() offset=0xF51 addr=0x987141
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::CNcbiApplication::x_TryMain(ncbi::EAppDiagStream, char const*, int*, bool*) offset=0x92A addr=0x1a9ddfa
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 ncbi::CNcbiApplication::AppMain(int, char const* const*, char const* const*, ncbi::EAppDiagStream, char const*, std::string const&) offset=0x773 addr=0x1a9c653
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 main offset=0x2CB addr=0x985f3b
/lib/x86_64-linux-gnu/libc.so.6 ???:0 __libc_start_main offset=0xF0 addr=0x7faa19d5e830
/analysis/Software/ncbi-blast-2.6.0+/bin/blastn ???:0 /analysis/Software/ncbi-blast-2.6.0+/bin/blastn() [0x985b29] offset=0x0 addr=0x985b29
Does anyone know how to solve it?
Looking for the reply! Thank you very much!
Myshu
I think the problem is with your query fasta. Did the error occur right away, or the BLAST run for a while before crashing? You can try to troubleshoot your input fasta by sppliting it in smaller chunks and testing then separately.
edit: see pointers at Question: ncbi error C++ exception.
The error occur right away. The result is null.
Yes, I tried to split the input fasta into small chunks...But even I split into 3000 chunks, and each of file only has about 600 sequences, the error still occurred....
Do you know what the largest number of input sequence is?
Select one (preferably short) sequence that causes the error and post it here.
Thanks for your reply. And I tried to select several sequences to do the test.
And I found that the sequence below causes the error...
But why? I can't find any different....
Hi 1291016966,
I tried with your sequence using the same command except for the database with blastn v2.6.0+. It worked well.
Hi 1291016966,
Did you solve the above problem? I got similar error when I run tblastn. In my case, the sequence below causes the error.
You might try to run the blast with a diff number of threads parameter (eg. 5 ) we've noticed that sometimes certain values of the num_threads are not working well, no idea why though.
Otherwise, perhaps activating/disabling low-complexity filter might help?
Thanks for your suggestion. I tried to run the blast with 5 threads and 10 threads. but I found a same error. I will try your other suggestion.
Hi nemo,
I got the following message: CFastaReader: Hyphens are invalid and will be ignored around line 2
Not sure but could it be because of the presence of the hyphens?
Excellent observation! Did not notice those
@ nemo : where do the hyphens come from? They should not be present in a fasta file.
Hi Thank you for your reply. hyphens come from our in-house developed program. yes, I noticed the messeage " CFastaReader: Hyphens are invalid and will be ignored around line 2". but this error is not the problem. I still have the same error (Error: NCBI C++ Exception:...) as 1291016966 mentioned above.
OK, so you did try when you removed the hyphens, right?
Did the formatting of your blastDB went ok? is there anything particular in the makeblastdb log file?
Are you able to blast anything with your 2.6.0 version?
Hi lieven,
yes, it seems there was no problem with makeblatdb. thank you for your suggestion. When I used old ver.(2.2.29), blast search was completed without any error!
Thank you,
I encountered the similar/same error message ("BLAST Database error: Error pre-fetching sequence data").
Once I used the newest? psiblast in .../ncbi-blast-2.10.0+/bin, the problem was solved.