Diamond with database of small kmers

[raeshrode]

Hello,

I am trying to run diamond blastp. I built my own diamond database with 9-12mer amino acids strings and I'm comparing those to .faa files with ~200aa. I have tried running diamond blastp with several different parameters and each time it results in 0 pairwise alignments. Below are some of the variations I've tried:

diamond blastp -d database.dmnd -q prac.faa -o match_prac.tsv

diamond blastp -d database.dmnd --matrix PAM30 -q prac.faa -o match_prac.tsv

diamond blastp -d database.dmnd --matrix PAM30 --ultra-sensitive -q prac.faa -o match_prac.tsv

diamond blastp -d database.dmnd --matrix PAM30 --ultra-sensitive --evalue 100 -q prac.faa -o match_prac.tsv

What do you suggest I do to improve my alignments? Is there a reason a database with short reads is unable to match?

Thanks, Rachel

[bbuchfink]

Diamond does not work well by default on very short sequences and needs to manual parameter tuning. I shared some tips here: https://github.com/bbuchfink/diamond/issues/832 https://github.com/bbuchfink/diamond/discussions/469 and in some more issues.