Hello,
I'm still rather new to blast+ and I'm trying to get an implemented tool to work locally(that already works somewhere else). The tool just completely skips blasting and doesn't even show me any error. It creates and fills all the required files perfectly, makes the database files and then tries to run the blastn module. But nothing happens. I just get a completely blank blast_output file. No errors in my error logs. So I tried to do some of the commands manually in the terminal.
Here is the makeblastdb command:
/home/student1/MGCV_liteTools/ncbi-blast-2.2.28+/bin/makeblastdb -in /var/www/localhost/blast_tmp/1386165610refseq_upload_file.fasta
-input_type fasta -max_file_sz 3500000000 -dbtype nucl -out /var/www/localhost/blast_tmp/1386165610_db
It seems to work fine, as I get the following result:
Building a new DB, current time: 12/04/2013 15:01:54
New DB name: /var/www/localhost/blast_tmp/1386165610_db
New DB title: /var/www/localhost/blast_tmp/1386165610refseq_upload_file.fasta
Sequence type: Nucleotide
Keep Linkouts: T
Keep MBits: T
Maximum file size: 3500000000B
Adding sequences from FASTA; added 7 sequences in 0.137179 seconds.
Then, I try to enter the blastn command:
/home/student1/MGCV_liteTools/ncbi-blast-2.2.28+/bin/blastn -db /var/www/localhost/blast_tmp/1386165610_db -query /var/www/localhost/blast_tmp/1386165610query_upload_file.fasta -evalue 0.001 -num_threads 4 -max_target_seqs 2000 -out /var/www/localhost/blast_tmp/1386165610blast.out -outfmt 5
The following error shows up:
Error: NCBI C++ Exception:
"/home/student1/MGCV_liteTools/ncbi-blast-2.2.28+-src/c++/src/corelib/ncbiobj.cpp", line 925: Critical: ncbi::CObject::ThrowNullPointerException() - Attempt to access NULL pointer.
Now I find it weird that I still get this error from the source map, even though I've deleted it already, and only kept the compiled version.
As I'm still very new to this, I tried some things to the best of my possibilities:
- validating the fasta's: both were correct;
- reinstall blast+: tried both the x32.tar.gz and the source one(and compiled it), still same error;
- tried changing num_threads to 1 and 2: didn't work.
Is anyone able to see what I'm doing wrong here?
UPDATE:
It still doesn't work, but atleast the tool is able to catch the error. Still no idea why it appears though.
Looks like you might have found a bug. I suggest you contact the NCBI BLAST folks via blast-help@ncbi.nlm.nih.gov