Using a VOM model for reconstructing potential coding regions in EST sequences (Q2271692)

The problem of recognition of protein coding DNA regions in expressed sequence tags is discussed. The authors propose algorithms based on variable order Markov (VOM) models for the distributions of coding and non-coding DNA sequences. In VOM models the order of dependence is not fixed but depends on the context (i.e., on the observed sequence). Results of experiments with real data sets are discussed. It is shown that the models are potentially efficient and robust to sequencing errors.