Question: Generate HMM profile without an MSA (hmmer)
0
gravatar for elenatea.russo
14 months ago by
Italy, Trieste, SISSA
elenatea.russo20 wrote:

Hello to everyone :)

I need to build a HMM profile with hmmer without an MSA, i.e., I have a set of protein sequences that align in couples but I don't have the multiple sequence alignment. As far as I understood, with hmmer suite you can create an hmmer profile from a msa or a msa from a hmmer profile. Since I have none of these, is there a way to build an msa alignment in hmmer without the profile? I would like to avoid the use of other programs.

Everything I could find about this situation is a slide from a "Tandy Warnow" lecture that says

Profile HMMs can be built from a given multiple sequence alignment – this is not too difficult. Profile HMMs can also be built from unaligned sequences. This is a bit complicated.

And no further explanation.

Am I missing something?

Bests Elena

hmmer msa hmm • 504 views
ADD COMMENTlink modified 14 months ago • written 14 months ago by elenatea.russo20

You want one profile to represent all the pairs alignments?

ADD REPLYlink written 14 months ago by Joe14k

0 gravatar for elenatea.russo just now by elenatea.russo • 0

Yes. But these are not uncorrelated alignmens, the situation is more like: A aligns with B that aligns with C that aligns with D etc... and they are local alignments, so I expect that a MSA should be possible to do. Thanks!

Elena

ADD REPLYlink written 14 months ago by elenatea.russo20

Well a few things spring to mind (though I can't speak personally to the 'more complicated approach') you've found a hint of above..

If A -> B -> C -> ... -> Z align to one another, there's no real reason you shouldn't be able to just do the MSA on all those sequences?

I'm less sure about this second one, but you might be able to have tools like CD-HIT emit an HMM for each of your clusters/pairs.

And, I think, you can make a profile HMM for each of your pairs (though on their own they won't be very good), and then you can build 'another layer' of HMMs off the first set. This would give you 1 HMM for all pairs.

Don't quote me on the last approach, but naively it seems like that might work.

ADD REPLYlink written 14 months ago by Joe14k

If A -> B -> C -> ... -> Z align to one another, there's no real reason you shouldn't be able to just do the MSA on all those sequences?

Yes, my problem is if it is doable with HMMER without the profile or if I have to use another software. I could use hmmer with your "last approach" but, yes, it looks a bit difficult to justify....

In any case, thanks :)

ADD REPLYlink written 14 months ago by elenatea.russo20

it is doable with HMMER without the profile

You've lost me a bit now. You want to use HMMER but without building a profile of your sequences?

ADD REPLYlink written 14 months ago by Joe14k

Sorry, you are correctly confused, i totally wrote the wrong thing.

*Yes, my problem is if it is doable with HMMER without the MSA *

My goal is to build the profile, indeed. So: can I build an HMM profile without an msa? (the answer should be: no). If not, can HMMER build the MSA or must I use another software? (As far as i've understood, no, i must use another software. correct?)

Bests

Elena

ADD REPLYlink modified 12 months ago • written 12 months ago by elenatea.russo20
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 2114 users visited in the last hour