Profile Hidden Markov Models

Question

[delcophantom8491]

What are the limitations of using the observed counts when deriving model parameters for profile HMM's?

[Jean-Karim Heriche]

This reads like an assignment. Don't expect people to do your homework for you.

Answer 1

[ashishbansal12300]

Profile HMMs are a model-based alternative to the ad-hoc multiple alignment approaches common for sequence family modeling. Since different multiple alignment software packages can yield very different alignments, use of the profile HMM approach is more robust than relying on a single multiple alignment to construct a consensus or a position-specific score matrix. While profile HMMs are widely used for protein family modeling, and databases of profile HMMs exist for protein sequence families, including the popular Pfam and SUPERFAMILY databases, profile HMMs have not generally been embraced as a tool for modeling DNA sequence families. Both of the major software packages for building and using profile HMMs explicitly focus on protein family modeling, though they retain some support for DNA. The Baum-Welch algorithm, which has problems with local optima in the context of protein sequences, is even more prone to premature convergence when there are four residues instead of twenty. Our hope is that the algorithmic contributions described here and in will open new possibilities for applications of profile HMMs to the DNA domain.

[Sej Modha]

Hi @ashishbansal12300,

It is not good practice to copy and paste answers from papers to this forum. If you think that a paper (e.g. here is the link to the paper you have posted the above text from), a website or another post answers the question then please refer to the original source of the information in your answer. It is not advisable to copy and paste text from other sources as answers on this forum.