[ssml] Redundancy in MSA for building HMM

[Dan Bolser]

On Wed, 6 Oct 2004, Goel, Manisha wrote:

>Hi All,
>
>I am trying to build Profile HMM for a couple of protein domain families
>(as defined in SCOP).
>For building a Multiple sequence alignment, I have picked up related
>members from PDB, but since the PDB often contains many structures of
>the same protein (with multiple ligands or so), my alignments have
>multiple copies of the same/highly similar sequences. 
> 
>Will the high redundancy of such an alignment effect the HMM model
>quality ?
>Since I will not be using these HMMs to look for distant relationships
>but only confine myself to studying the members within a family .. I was
>assuming this should not be problem.
>Any suggestions ?

Why do you want to use the HMM in this case?

You could try using hmmer without calibrateing the modle. 


> 
>Thanks,
>-Manisha
>