Diamond runtime + optimisation

corend

New member
Hello,

I am running diamond blastx on a set of ~10k sequences of ~1kb against nr.

I have a server with 24CPU, and 128 Go of memory.

Which parameters can I use to optimize the runtime of diamond?

Is there an option to use to retrieve only the best hit for each sequence? Does it take less time than retrieving all the hits?

Do you have an idea of the expected runtime?

Here is the command line I am running at the moment:
Bash:
nohup diamond blastx -d /database/nr -q /data/query.fa -o output.bl -f 6 qseqid sseqid stitle -e 1e-10 -k 1 -p 24 >& blast.log &
Thanks a lot,
 

Benjamin Buchfink

Administrator
Staff member
You can use -b5 -c1, that will help. Using -k1 will get the best hit only, and it's definitely faster than retrieving all hits.

I'm not sure about the expected runtime. Diamond generally is more efficient if you use large query files (>1 million sequences).
 
Top