Predicting Human Nucleosome Occupancy from Primary Sequence

• DOI: 10.1371/journal.pcbi.1000134
• Publication Date: 2008-08-21T18:08:55Z
• Authors: Shobhit Gupta; Jonathan Dennis; Robert E. Thurman; Robert Kingston; John A. Stamatoya...; William Stafford ...
• Keywords: Micrococcal nuclease; ENCODE

Nucleosomes are the fundamental repeating unit of chromatin and comprise structural building blocks living eukaryotic genome. Micrococcal nuclease (MNase) has long been used to delineate nucleosomal organization. Microarray-based nucleosome mapping experiments in yeast have revealed regularly-spaced translational phasing nucleosomes. These data train computational models sequence-directed nuclesosome positioning, which identified ubiquitous strong intrinsic positioning signals. Here, we successfully apply this approach from human chromatin. The predictions made by human-trained yeast-trained strongly correlated, suggesting a shared mechanism for sequence-based determination occupancy. In addition, observed striking complementarity between classifiers trained on experimental weakly versus heavily digested MNase samples. former case, resulting model accurately identifies nucleosome-forming sequences; latter, classifier excels at identifying nucleosome-free regions. Using able identify several characteristics nucleosome-disfavoring sequences. First, combining results each applied de novo across ENCODE regions, reveals distinct sequence composition periodicity features Short runs dinucleotide repeat appear as hallmark sequences, while sequences contain short periodic GC base pairs. Second, show that is most frequently predicted flanking suggest major vivo boundary-event-driven affirm classical statistical theory