“…These features, which can be considered controls, included six that were previously shown to be nonrandomly associated with UCE positions: copy number variants (CNVs), cancer-specific copy number alterations (CNAs), genes, exons, introns, and segmental duplications (SDs) (Chiang et al, 2008; Derti et al, 2006; McCole et al, 2014). They also included open chromatin, since UCEs have been linked to transcriptional activity (reviewed in Baira et al, 2008; Fabris and Calin, 2017; Harmston et al, 2013), repetitive elements, which UCEs avoid (Bejerano et al, 2004; Chiang et al, 2008; Derti et al, 2006; McCole et al, 2014), and GC content, which is associated with the positions of CNVs (Koren et al, 2012). We divided the genome into equally sized bins and, because domains and the nine control features span a vast range of sizes, our analyses involved multiple iterations using a range of bin sizes (20, 50, and 100 kb).…”