Table 4.

Correlations across five datasets taking segmentation into account.

Hyman	0%	25%	50%	75%	100%	SD	MAD
Both Unsegmented	−0.859	−0.057	0.159	0.373	0.956	0.303	0.246
Expression Segmented	−0.805	−0.025	0.211	0.432	0.974	0.317	0.259
Copynumber Segmented	−0.902	−0.022	0.219	0.453	0.969	0.324	0.267
Both Segmented	−0.734	0.259	0.471	0.652	0.976	0.281	0.224
Heidenblad	0%	25%	50%	75%	100%	SD	MAD
Both Unsegmented	−0.777	−0.051	0.147	0.34	0.949	0.276	0.224
Expression Segmented	−0.746	0.069	0.268	0.463	0.948	0.285	0.229
Copynumber Segmented	−0.804	−0.008	0.196	0.395	0.952	0.288	0.234
Both Segmented	−0.691	0.247	0.464	0.645	0.975	0.288	0.231
Pollack	0%	25%	50%	75%	100%	SD	MAD
Both Unsegmented	−0.547	−0.028	0.106	0.248	0.898	0.208	0.165
Expression Segmented	−0.607	0.082	0.241	0.399	0.947	0.226	0.183
Copynumber Segmented	−0.54	0.005	0.166	0.331	0.914	0.231	0.186
Both Segmented	−0.396	0.327	0.479	0.62	0.971	0.214	0.172
Zhao	0%	25%	50%	75%	100%	SD	MAD
Both Unsegmented	−0.946	−0.127	0.196	0.488	0.989	0.406	0.338
Expression Segmented	−0.939	0.085	0.42	0.675	0.99	0.401	0.331
Copynumber Segmented	−0.968	−0.115	0.227	0.525	0.99	0.418	0.348
Both Segmented	−0.923	0.309	0.623	0.819	0.997	0.383	0.304
Kim	0%	25%	50%	75%	100%	SD	MAD
Both Unsegmented	−0.803	0.051	0.241	0.42	0.972	0.263	0.214
Expression Segmented	−0.726	0.124	0.328	0.522	0.932	0.279	0.228
Copynumber Segmented	−0.722	0.077	0.287	0.479	0.957	0.278	0.228
Both Segmented	−0.644	0.275	0.482	0.675	0.924	0.282	0.228

Notes: SD represents standard deviation; MAD represents mean absolute deviation. Both Unsegmented refers to correlation coefficients between gene expression and copy number based on the raw expression number value, after taking the preprocessing steps described in the paper. Expression Segmented refers to running the algorithm of Olshen et al. (2004) on the gene expression data on individual samples and to then calculate the correlation coefficients between gene expression and copy number expression. Copynumber Segmented refers to running the algorithm of Olshen et al. (2004) on the copy number expression data on individual samples and to then calculate the correlation coefficients between gene expression and copy number expression. Both Segmented refers to running the algorithm of Olshen et al. (2004) on the copy number and gene expression data on individual samples and to then calculate the correlation coefficients between gene expression and copy number expression.