Set database size in MMseqs2 (like -dbsize in blast)

Question

[duncan.sussfeld]

Hello all,

I would like to search a sequence database using MMseqs2. Because the database might change in size during my analyses, I would like to be able to specify a fixed database size to avoid creating inconsistencies in reported E-values.

I know BLAST and DIAMOND allow this with the -dbsize option, but I cannot find anything doing this in MMseqs2 documentation. Is there by any chance an option I'm missing?

Thanks a lot!

Answer 1

[duncan.sussfeld]

For information, when enquiring about this on the MMseqs2 GitHub page, I got the following response from a member of the development team:

Currently we don’t have an option like that. However, I encountered myself multiple cases where I could have used this feature. I can take a look into implementing this sometimes soonish (a bit busy right now though).