Prediction of regulatory elements in mammalian genomes using chromatin signatures

摘要: Recent genomic scale survey of epigenetic states in the mammalian genomes has shown that promoters and enhancers are correlated with distinct chromatin signatures, providing a pragmatic way for systematic mapping these regulatory elements genome. With rapid accumulation modification profiles genome various organisms cell types, this based approach promises to uncover many new elements, but computational methods effectively extract information from datasets still limited. We present here supervised learning method predict on their unique signatures. trained Hidden Markov models (HMMs) histone data known enhancers, then used HMMs identify promoter or enhancer like sequences human Using simulated annealing (SA) procedure, we searched most informative combination optimal window size marks. Compared previous methods, HMM can capture complex patterns modifications particularly weak signals. Cross validation scanning ENCODE regions showed our outperforms profile-based enhancers. also including more marks further boost performance method. This observation suggests is robust capable integrating multiple To demonstrate usefulness method, applied it analyzing wide ChIP-Seq three mouse lines correctly predicted active inactive positive predictive values than 80%. The software available at http://http:/nash.ucsd.edu/chromatin.tar.gz .