Open rmostowy opened 5 years ago
Does it also crash if you run the same search with HHsearch instead?
No it doesn't, nor does it crash with the command above and -n 1
option.
Will your database remain small (<100k profiles)? We generally recommend using HHsearch for small databases due its slightly higher sensitivity. Also iterations don't make sense to search against a small database. Generally the workflow is to build a highly diverse query profile by searching against a large profile database (Uniclust) and then doing a HHsearch profile-profile search for maximum sensitivity.
I'll try to take a look at the issue anyway in the next few days. Thanks for the bug report!
Thank you Milot, that is quite helpful as I've only just started using the HHSuite. I'll reconsider my approach. Indeed, the example I used is based on a very small input, however eventually I would like to extend it to a large protein database.
One more question. Can one generate the 'msa' file, which is the output of the mmseqs result2msa
command, with custom MSA fasta files? Basically, I cannot find anywhere steps to go from a number of custom MSAs:
data/msa0001.aln
data/msa0002.aln
.
.
.
etc., to a custom HHSuite database with the hhm, cs219 and c23m files. I searched for this extensively but didn't find any examples.
Hey @rmostowy , just a quick question about #3 in your steps, as I am hung on that step. I made my MSA using muscle and so all I have is one file as output (fa.out file). ffindex_apply in step #3 looks like it requires two files msa_ca3m.ff{data,index}, where did you get the index file from? Does mmseqs give you this? thanks for your help.
Expected Behavior
I am trying to build a custom database from my own fasta file by (1) clustering all protein sequences using
mmseqs
and then (2) following the steps of [https://github.com/soedinglab/hh-suite/wiki#building-customized-databases], and then search this database with a query sequence.Current Behavior
While clustering and database creation finish without problems, I run into problems when performing multiple iterations of
hhblits
search against the newly build database.Steps to Reproduce (for bugs)
All files mentioned here are compressed in the attached files.zip.
I begin with a fasta file
input.fa
, and then execute the following series of commands:After this, I perform a search with a query sequence using
HH-suite Output (for bugs)
The full output of the query search is:
Context
The context is that I would like to create a matrix of HMM-HMM similarities for a numer of different protein sub-families, each of which has an individual HMM profile. Hence, I want to combine all HMMs into a single database to search against.
Your Environment
Include as many relevant details about the environment you experienced the issue in.