An R package for the analysis of array-based aCGH data using non-homogeneous Hidden Markov Models with Reversible Jump Markov Chain Monte Carlo.
The method implemented in RJaCGH allows us to incorporate distance between probes/clones (the non-homogenous in non-homogeneous HMM). We return posterior probabilities that a gene/region is altered, using Markov Chain Monte Carlo. The method does not require you to pre-specify the number of states, as we use Reversible Jump for transdimensional moves. Finally, we use Bayesian Model Averaging for incorporating model uncertainty.
The details of our method are provided in a COBRA Preprint Series article http://
RJaCGH is part of the Asterias project, and will soon be incorporated to its web-based tools.