Diamond sensitivity setting for long reads

User Questions
1

Hi,

I noticed that Diamond has now several sensitivity modes (3. Command line options · bbuchfink/diamond Wiki · GitHub). I was wondering if you have a recommendation which mode might work best when assigning long sequences (assembled PacBio reads) taxonomically in MEGAN. The default mode is designed for finding hits of >60% identity and short read alignment.

I have been using the following command so far:
diamond blastx --db database.dmnd --query assembly.fasta --more-sensitive --frameshift 15 --range-culling --top 10 --outfmt 100 --out assembly.daa

Thank you very much in advance!
Anjuli

2

It’s a tradeoff of sensitivity vs running time. While more sensitive is better, how long are you will to wait for DIAMOND to complete? As you request more sensitivity from DIAMOND, the sensitivity will increase a little, but the runtime will increase a lot.

1 Like