This feature is still in development... see Ticket:143
The algorithm is as follows:
- Use a pairwise stemloc alignment to seed a Covariance Model-style Stochastic Context Free Grammar in DART.
- The Covariance Model is then optimized by Expectation Maximization.
- The Cocke Younger Kasami alignment is reported.
The concept of using Expectation Maximization to train profiles of repeated motifs and do alignment at the same time was used for HMMs in MEME (Bailey & Elkan: The value of prior knowledge in discovering motifs with MEME. Proc Int Conf Intell Syst Mol Biol 1995;3:21-9.) and implicitly in SAM (Brown et al.: Using Dirichlet mixture priors to derive hidden Markov models for protein families. Proc Int Conf Intell Syst Mol Biol 1993;1:47-55.).
-- Ian Holmes - 03 Apr 2005