[Biojava-l] single and double headed HMM's

Matthew Pocock mrp@sanger.ac.uk
Mon, 22 Jan 2001 11:14:13 +0000


Mark Schreiber wrote:

> Hi,
> I am fairly familiar with HMM's but I am not shure what is meant by the
> term double headed. Can someone explain??


A single-head HMM emits (or aligns) a single sequence with a set of states.
A two-head HMM emits (or aligns) two sequences with a set of states. An
n-head HMM emits (or aligns) n sequences with a set of states. However many
heads the model has, it emits paralell lists of symbols, one line for each
sequence being aligned, with gaps used to space the sequences out to the
same length.

A traditional one-head model would be a trans-membrain model, a profile HMM
or a MEME weight-matrix. A traditional example of a two-head model would be
pair-wise alignment, or aligning protein to genomic sequences. MSA performs
n-head alignment to give you the (nearly) optimal alignment of n sequences.
Each head may have a distinct alphabet - you could align DNA and PROTEIN
and cDNA if you wanted - if you could think of something usefull that this
did. An obvious and fun 3 head algorithm would be to align genomic to
genomic via cDNA or PROTEIN to make a comparative gene finder.

Should this go into the docs somewhere?

> Mark
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Mark Schreiber                  Ph: 64 3 4797875
> Rm 218                          email mark_s@sanger.otago.ac.nz
> Department of Biochemistry      email m.schreiber@clear.net.nz
> University of Otago
> PO Box 56
> Dunedin
> New Zealand
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l