Blastp giving empty output file

Question

[pes]

I am very new to blasting. I am using the terminal to do a blastp, it starts running but has taken up to 2 days and the output file is also empty. I have read somewhere that we can't see anything from the output file until blast has stopped and that's why I have used commands like watch and list to see if anything gets added to this file but those are also blank!

I would appreciate any help!

my script: pwd/ncbi-blast-2.6.0+/bin/blastp -query XXX -db XXX -outfmt 6 -out XXX.txt

[pes]

Thanks for answering! the size of the query I am using is 53680861, nothing is beind added to output file even when I run ls -l there is a zero. I will try running with more threads

[shenwei356]

PLEASE click ADD COMMENT below answers to respond.

[pes]

Thanks for answering! the size of the query I am using is 53680861, nothing is being added to output file even when I run ls -l there is a zero. I will try running with more threads when I look at top I am seeing it running!

[GenoMax]

Please use ADD COMMENT/ADD REPLY when responding to existing posts to keep threads logically organized.

Do not use SUBMIT ANSWER to add new information for original question.

Answer 1

[shenwei356]

These are lots of factors affecting the computation time:

• BLAST database size. e.g., NCBI NR is very very big.
• Query sequences size.
• Parameters.
  • -num_threads. You did not set -num_threads, which defines number of CPUs to use. I think this the main cause for your case.
• Computer hardware, including CPU, RAM, HDD and so on.

And I suggest your save result in -outfmt 11 and then convert it the other formats, in case you want more information than these in -outfmt 6.

[pes]

Thank you for your help!

Answer 2

[Jake Warner]

What is that number? Surely you're not blasting 53 million sequences? If so, and you are blasting against nr, with the one thread above I would guess it will take a few years to run.

Could you provide the results to:

grep -c '^>' query.fasta

To answer your question, if I remember correctly BLAST+ buffers the results stream so at some point ~100 results should appear in out.txt. In addition to the -num_threads parameter you can also set the word size higher eg: -word_size 5

[pes]

Thank you for your help!

Answer 3

[328558608]

first，check your query file type is protein sequence and your db index is protein type; In ordinary, you could see the result adding to the out file when blast+ is running. Make sure your command is correct and the command is in system background during 2 days.

[pes]

Thank you for your help!