Generative/discriminative models for nucleosome positioning

DNA of eukaryotic cells is organized into repeating nucleosomes. As nucleosomes significantly limit the accessibility of their DNA to transcription factors and other DNA-binding proteins, their positions play an essential role in regulations of gene activities [4, 8]. Recent genome wide experiments indicate that DNA sequences themselves strongly influence nucleosome positioning by enhancing or reducing their binding affinity to nucleosomes, therefore providing an intrinsic cell regulatory mechanism. Some sequence features are known to be nucleosome forming or nucleosome inhibiting. In this paper, we propose a computationally efficient method for nucleosome positioning. By using generative/discriminative models, the proposed methods can categorize nucleosome forming and nucleosome inhibiting local DNA sequences in a much lower dimensional feature space. Experiment results shows our methods have better performance than discriminative methods.