Building customized database using HHblits

Question

[siu]

Hi,

Sorry for asking this naive question

I have 7000 FASTA sequences (not MSA) and I want to build a customized database of these and search against themselves using HHblits. I am following HH-suite tutorial (https://github.com/soedinglab/hh-suite/wiki#building-customized-databases) but I am getting an error every time which is "Input file is missing" in ffindex_apply step. I have posted this on their GitHub page as well but I didn't got any response. I want to ask if someone has recently run HHblits from HHsuite (I have tried several times but not able to do it)? What are the exact steps that I need to run to find the homologous sequences of these 7000 FASTA sequences against themselves?

It will be a great help

Thanks in advance

[Dunois]

Is there any reason why you need to use HHblits specifically here?

[siu]

Actually, I want to use different tools like Diamond, MMseq2, Jackhmmer, HHblits (different principles of finding homologs) and want to rank the homologs based on aggregated score. I want to find the true/remote homologs by this criteria (Used by this paper, https://www.nature.com/articles/srep32333#Sec6) If I can have any tool better than HHblits in terms of speed and accuracy that would also be fine for me.

Answer 1

[Mensur Dlakic]

There are many ways to build a HHblits database, but the way described in the link should work. Many people probably have their own pipelines - myself included - but that would take too long to explain and still probably wouldn't be any simpler.

A starting point for doing anything is being able to run HHblits, so you may try your luck with that first. Note that there are many dependencies and databases needed, and you will need to provide a more detailed explanation about your problem rather than just saying that HHblits is not working. There are also posts on the right-hand side of this page that could be useful.