Figure 5

Nucleosome organization in exons and introns. A) Barplot of exon and intron nucleosome occupancies. The mean normalized nucleosome occupancy in introns and exons for each dataset is plotted. Error bars show the standard error. B) Barplot of predicted exon and intron nucleosome occupancies. Same as (A), but using normalized nucleosome occupancy profiles predicted by N = 2 position-independent models fit on the indicated datasets. C) In vitro nucleosome occupancy in exons and introns grouped by GC content. Exons and introns were divided into three equally sized groups of high, medium, and low GC content. Introns were aligned on their center, and exons were aligned to their 3’ ends (left) and 5’ ends (right). Mean normalized nucleosome occupancy in each group is plotted against the distance from the center for introns, and the distance from either the 3’ or the 5’ boundary for exons. Averages x bases upstream of the 3’ boundary or downstream of the 5’ boundary are calculated only among exons of length ≥ x. The average intron nucleosome occupancy a distance x from the intron center is calculated only among introns of length ≥ 2x. Dashed curves show standard errors of the mean. The nucleosome occupancy profile is from the Rsa I in vitro assay. D) In vivo nucleosome occupancy in exons and introns grouped by GC content. Same as (C), but using in vivo data from Valouev et al. [39]. E) Exon and intron nucleosome occupancy grouped by gene expression levels. Same as (A), except that exons and introns are from the genes with high and low expression levels. Expression levels were inferred from SAGE data. All tagged genes were ranked by the abundance of SAGE tags, with high and low expression groups corresponding to the top and bottom 10%, respectively (see Methods).