Hi @xiezhq ,
the phmmer based lookup of single AA is responsible for a large part of the overall runtime. As phmmer internally more or less behaves like a normal blastp, it might be worth to replace phmmer by a faster solution like:
https://github.com/bbuchfink/diamondhttp://www.diamondsearch.org/index.php
For this small protein database clusters.single.faa the homology search of all AA seqs of a Pseudomonas aeruginosa genome took less than 1 sec and only 18 Mb RAM with 2 cores:
time -v diamond blastp --query paeruginosa.faa --db clusters.single.dmnd --threads 2 --out diamond.tsv --outfmt 6 --header"
User time (seconds): 1.03
System time (seconds): 0.03
Percent of CPU this job got: 181%
Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.58
Maximum resident set size (kbytes): 18624
Are there any good counter reasons? Otherwise this might significantly reduce overall runtimes.
Best regards
Hi @xiezhq , the
phmmer
based lookup of single AA is responsible for a large part of the overall runtime. Asphmmer
internally more or less behaves like a normalblastp
, it might be worth to replacephmmer
by a faster solution like: https://github.com/bbuchfink/diamond http://www.diamondsearch.org/index.phpFor this small protein database
clusters.single.faa
the homology search of allAA
seqs of aPseudomonas aeruginosa
genome took less than 1 sec and only 18 Mb RAM with 2 cores:Are there any good counter reasons? Otherwise this might significantly reduce overall runtimes. Best regards