Association between histone modifications and retroviral integrations in CD34⁺ HPCs. Histone modifications around (± 1000 bp) a subset of retroviral and random integration sites (ISs) were evaluated by ChIP-on-chip technology. (A) Chromatin immunoprecipitated from cytokine-stimulated CD34⁺ HPCs was amplified, fluorescently labeled, and hybridized to custom-designed “integrome” arrays, where we spotted tiled oligonucleotides (50-60mers) covering 1000 bp upstream and downstream of each insertion site. (B) For each experiment, chromatin immunoprecipitated with the antibody of interest (sample) was cohybridized with an equal amount of chromatin immunoprecipitated with agarose beads only and labeled with a different fluorophore (total input). ChIP peaks were statistically defined starting from sample-to-input raw signal ratios (the threshold for peak definition, set at the 95th percentile of the dataset distribution, is specified for each sample, and indicated by a dashed horizontal line). A representative output of ChIP-on-chip experiments is given for a subset of antibodies (against H2BK5me1, H3K27me3, H3K4me3, and H3K9ac) around 3 MLV integration sites targeting the NF1/EVI2A/B locus on chromosome 17. (C) Percentage of MLV, HIV, and random ISs with ≥ 1 peak of the specified histone modifications in the flanking ± 1-kb region. Epigenetic marks are grouped according to the genomic region they are classically associated to (group A, enhancers and promoters of active genes; group B, promoters and gene bodies of actively transcribed genes; group C, inactive genes and heterochromatic regions). Burgundy, red, and yellow sections inside group-A MLV bars indicate the relative proportion of TSS-proximal, intragenic, and intergenic integrations to the observed enrichment. Asterisks denote the level of overrepresentation or underrepresentation of MLV and HIV ISs with respect to random sites: **P < .005, ***P < .0005 by 2-sample test for equality of proportions with continuity correction.