Once relevant TF binding profiles are identified, the challenge shifts to the identification of reliable TFBS within the broader ChIP-Seq regions.
These overlaps do not however reflect some kind of broad ChIP-chip artifact.
Broad ChIP-Seq data does not generally produce clearly delimited regions; hence, one cannot speak of peaks.
The flexibility of Pyicos is further demonstrated by applying the EA protocol to broad ChIP-Seq data.
Furthermore, we illustrate the flexibility of Pyicos by showing its effectiveness on other HTS datatypes: broad ChIP-Seq signals and CLIP-Seq data.
We next provide evidence that Pyicos EA approach is also suitable for the analysis of broad ChIP-Seq data when comparing two conditions.
We performed EA on broad ChIP-Seq signals using ENCODE data for H3K36me3 and RNAPII, and correlated the results to the enrichment of RNA-Seq data, for the K562 and NHEK cells (ENCODE Consortium, 2011).
Thus, the EA of Pyicos using RPKM provides a tool to analyse broad ChIP-Seq data of various sorts, which would otherwise require a combination of approaches to be analysed (Young et al., 2011).
However, in HSF-D and HSF-E, the CENP-A binding domains are partially overlapping in the horse genome sequence and correspond to the left and the right part of the broad ChIp-on-chip peak, respectively; the overlapping region roughly corresponds to the centre of the broad peak.
ChIP-seq experiments reveal extensive chromatin modification within the 263-kb regincludinguding H3K4me1, H3K4me2, H3K4me3, H3K9ac, H3K27me3 and H3K27ac (http://genome.ucsc.edu/ ENCODE Histone Mods, Broad ChIP-seq; Fig. 3d), as would be expected for a region containing several protein-coding genes.
