Mining biologically active patterns in metabolic pathways using microarray expression profiles

摘要: We present a new probabilistic framework for analyzing metabolic pathway with microarray expression profiles. Our purpose is to find biologically significant paths and patterns in given pathway. approach first builds Markov model using graph structure of known pathway, then estimates parameters mixture the models data, based on an EM algorithm. In our experiments, we used main glycolysis evaluate effectiveness method. measured performance method comparing that another method, supervised learning manner, found significantly outperformed which was trained by data only. further analyzed obtained number biological findings frequent (paths) long-range correlations