tileHMM-package: Hidden Markov Models for ChIP-on-Chip Analysis

In tileHMM: Hidden Markov Models for ChIP-on-Chip Analysis

Description

This package provides methods and classes to build HMMs that are suitable for the analysis of ChIP-on-chip data. The provided parameter estimation methods include the Baum-Welch algorithm and Viterbi training as well as a combination of both. The latter provides results identical to the Baum-Welch algorithm but is conciderably faster.

Details

Package:	tileHMM
Type:	Package
Version:	1.0-5
Date:	2012-03-19
License:	GPL (>= 2)

Hidden Markov models are represented as objects of class hmm or derived classes. Function getHMM provides an easy to use interface to create contHMM objects with emission distributions of class tDist from a set of parameters. Function hmm.setup can be used to create HMMs with initial parameter estimates obtained from data.

To optimise initial parameter estimates Viterbi training and the Baum-Welch algorithm are provided by this package. Function viterbiEM provides a convenient way to use a combination of both methods.

Author(s)

Peter Humburg

Maintainer: Peter Humburg [e-mail]

References

Humburg, P. and Bulger, D. and Stone, G. Parameter estimation for robust HMM analysis of ChIP-chip data. BMC Bioinformatics 2008, 9:343.

See Also

Classes provided by this package: hmm, contHMM, dist, discDist, contDist, tDist

Important functions: viterbiEM, baumWelch, viterbiTraining

tileHMM documentation built on May 30, 2017, 3:41 a.m.