Genome-wide analysis of DNA repetitive elements in estrogen receptor binding sites.

Estrogen Receptor alpha (ERα) is one of the key molecular players in most human breast cancers and the influence of ERα at the genomic level in response to ligands has been widely described (Carroll 2006). Unliganded ERα was also recently correlated to epithelial-mesenchymal transition genes (Cardamone et al., 2009). Several studies showed that DNA repeats in the human genome are significantly associated with transcription factor binding sites (Bourque 2008, Kunarso 2010). Here, we first set out to examine the binding of ERα to DNA using several chromatin immunoprecipitation sequencing (Chip-Seq) public datasets on MCF7 and T47D cell lines. We then performed a genome-wide bioinformatics analysis of DNA repetitive features and Transcription Factor Binding Sites (TFBSs) considering ERα binding sites obtained in different conditions. In particular, we have set up a bioinformatic pipeline that is able to explore the behavior of genomic repeat features mapped on Chip-Seq binding peaks and to assess the statistical enrichment of single repeats (or class of repeats) by means of rigorous Monte-Carlo simulations, in the human genome. Furthermore, we have performed a genome-wide scan for putative TFBSs overlapping Chip-seq peaks and repetitive regions, employing canonical Positional Weight Matrices (PWMs) reported in the Jaspar database (http://jaspar.genereg.net/) and using affinity scores assigned using a standard Log-Likelihood Ratio scoring function with background frequencies evaluated on human intergenic sequences. A few repetitive sequences and TFBSs showed a remarkable enrichment in ERα sites thus proposing a putative mechanism for the definition of estrogen receptor binding events.