Get cluster sequence identities about mmseqs2 HOT 4 CLOSED

Hi Antonio,

You can do something like this:

mmseqs cluster db clu tmp
mmseqs align db db clu aln -a
mmseqs createtsv  db db aln aln.tsv

The -a parameter computes the whole backtrace, you can also use --alignment-mode 3, but that computes the backtrace anyway since it needs it for the seq. id.
--alignment-mode 1 and 2 only approximate the sequence identity.

Thank you for the feedback! We will add this to the FAQ and also discuss if it make sense to provide the same output also for Uniclust in future releases.

from mmseqs2.

Thank you very much! I will give it a try tonight and be back to you.

We will add this to the FAQ and also discuss if it make sense to provide the same output also for Uniclust in future releases.

In our case, this information will help us to evaluate the quality of the clusters.

Cheers
Antonio

from mmseqs2.

One more possible problem: the default --max-seqs is too low for this task. You should set it to some very large number (--max-seqs 2147483647). So you will get every sequence identity of every cluster member.

from mmseqs2.

Hi Milot,
I had a quick look to the results and seems that it work perfectly. I will have a more thorough look to the results in case we are missing any pair.
Many thanks

Cheers
Antonio

from mmseqs2.