Cells were crosslinked and sonicated followed by chromatin immunoprecipitation as previously described (29 (link)). Antibodies against H3K9ac, H3K18ac, H3K27ac, H4K8ac and H3K27me3 were obtained from Abcam (ab4441, ab1191, ab4729, ab15823 and ab6002), RXRα, and p300 from Santa Cruz (sc-553x and sc-32758x). DNA was purified using the MinElute Spin Columns Kit (Qiagen) and input chromatin DNA was used as control. Purified DNA was sequenced by the McGill University Genome Quebec Innovation Centre with Illumina HiSeq 2000. Sequencing reads were mapped to the mouse genome build mm9 using Bowtie, allowing for three mismatches and reporting the single best alignment per 50 bp read. Picard was used to filter out replicated reads (http://picard.sourceforge.net/), and BAM files were converted into BED files with the BEDTools suite (33 (link)). For visualization of ChIP-seq signals in the Integrative Genomics Viewer, aligned reads were extended by 125 bp at their 3′ end and basewise signal intensity was computed. Local peaks in read enrichment were identified using MACS (34 (link)) (v1.0.0) with a P-value threshold of 1 × 10−5.