Table 3. UCEs are depleted from pooled ^{de nov}°CNVs, enriched in pooled ^cancerCNAs, and depleted from pooled ^somaticCNVs and high passage ^iPSCNVs.

	Observed overlap		Expected overlap (bp)		Result
Dataset	Number of UCEs	bp	Mean	Standard deviation	Proportion	P value	obs/exp	Outcome
				de nov° CNVs
Pooled de nov° CNVs a	4	967	2447	866	0.024	0.044	0.395	Depleted
				cancerCNAs
Pooled cancerCNAs b	642	171554	157016	3963	0.000	1.2×10−4	1.093	Enriched
TCGARN 2008	60	15670	16549	2091	0.345	0.337	0.947	Neither
Walter 2009	121	32312	29544	2652	0.143	0.148	1.094	Neither
Beroukhim 2010	172	46060	47042	3270	0.383	0.382	0.979	Neither
Taylor 2010	84	23669	19685	2384	0.058	0.047	1.202	Neither
TCGARN 2011	259	67447	57988	3689	0.005	0.005	1.163	Enriched
Curtis 2012	51	13893	11872	1869	0.141	0.140	1.170	Neither
TCGARN 2012 breast	156	42421	26852	2677	0.000	3.0×10−9	1.580	Enriched
TCGARN 2012 colon	26	6813	10016	1672	0.021	0.028	0.680	Neither
TCGARN 2012 squamous	218	58477	51127	3424	0.020	0.016	1.144	Enriched
Robinson 2012	60	16125	11569	1844	0.010	0.007	1.394	Enriched
Walker 2012	893	240548	233338	1433	0.000	2.4×10−7	1.031	Enriched
Zhang 2012	37	10028	12325	1934	0.120	0.118	0.814	Neither
TCGARN 2013	83	21061	28531	2709	0.001	0.003	0.738	Depleted
				somaticCNVs
Pooled somaticCNVs c	487	130694	142562	4019	0.004	0.002	0.917	Depleted
Forsberg 2012	23	6576	4195	1061	0.018	0.012	1.568	Enriched
Jacobs 2012	265	69836	70178	3831	0.459	0.464	0.995	Neither
Laurie 2012	264	69935	67605	3559	0.259	0.256	1.034	Neither
McConnell 2013	221	60353	70021	3747	0.005	0.005	0.862	Depleted
				iPSCNVs
Pooled low passage iPSCNVs d	15	3522	3235	1001	0.358	0.387	1.089	Neither
Pooled medium passage iPSCNVs d	13	3633	6006	1283	0.020	0.032	0.605	Neither
Pooled high passage iPSCNVs d	6	1279	3912	1033	0.003	0.005	0.327	Depleted
Hussein 2011 low passage	13	2926	3086	943	0.466	0.433	0.948	Neither
Hussein 2011 medium passage	12	3428	5196	1240	0.069	0.077	0.660	Neither
Hussein 2011 high passage	1	211	1980	761	0.001	0.010	0.107	Depleted
Laurent 2011 high passage	5	1068	1962	768	0.108	NA	0.544	NA

Here, we show the relationship between 896 HMR-HDM-HC UCEs and ^{de nov}°CNVs, ^cancerCNAs, ^somaticCNVs, and ^iPSCNVs, reporting the results for pooled datasets as well as all individual datasets that met our requirement for 20 Mb of coverage (Table S3). Individual CNV and CNA datasets are named according to the first author and the year of the study.

^aThe pooled ^{de nov}°CNV dataset included datasets from Xu 2008 [69], Itsara 2010 [70], Malhotra 2011 [71], and Sanders 2011 [72], which were too small to be considered on their own.

^bThe pooled ^cancerCNA dataset included all the ^cancerCNA datasets listed in this table, except for Walker 2012 [87], which was excluded to avoid bias from its extensive coverage of the genome, and also included the datasets Bullinger 2010 [78], Nik-Zainal 2012 [85], Holmfeldt 2013 [89], and Weischenfeldt 2013 [91], which were too small to be considered on their own.

^cThe pooled ^somaticCNV dataset included the four ^somaticCNV datasets listed in this table as well as Piotrowski 2008 [63] and O'Huallachain 2012 [67], which were too small to be considered on their own.

^dThe pooled ^iPSCNV datasets were comprised of CNVs from low, medium, and high passage iPS cells from the two datasets Hussein 2011 [100] and Laurent 2011 [98]. Proportion, P-value, and obs/exp, as described for Table 1. Outcome: determined with a one-tailed test (α = 0.05) for the pooled ^{de nov}°CNV dataset because dataset was analyzed prior to our discovery that CNVs can be enriched for UCEs; all other assessments of depletion or enrichment carried out with a two-tailed test (P≤0.025 in each tail for an overall α of 0.05). NA (not applicable): expected overlaps not normally distributed, precluding a Z-test.