Table 1.

Benchmark Pearson correlation coefficient of ACDA/CDA models on select tissues and studies

Dataset	Tissue	Train	Test	ACDA	CDA	EN	EN-ACDA	ACDA-EN-ACDA
GDSC1&2	Breast	127	64	0.439 ± 0.080	0.377 ± 0.066	0.394 ± 0.101	0.492 ± 0.083	0.520 ± 0.083
GDSC1	Breast	79	40	0.596 ± 0.032	0.529 ± 0.033	0.700 ± 0.063	0.733 ± 0.032	0.717 ± 0.200
GDSC2	Breast	108	55	0.706 ± 0.054	0.449 ± 0.038	−0.079 ± 0.021	0.594 ± 0.106	0.675 ± 0.064
GDSC1&2	Bladder	73	37	0.501 ± 0.013	0.481 ± 0.037	0.597 ± 0.039	0.656 ± 0.025	0.613 ± 0.015
GDSC1	Bladder	60	30	0.602 ± 0.042	0.634 ± 0.037	0.628 ± 0.040	0.658 ± 0.040	0.647 ± 0.041
GDSC2	Bladder	58	29	0.098 ± 0.083	0.183 ± 0.061	0.454 ± 0.072	0.436 ± 0.061	0.318 ± 0.058
DC-AZ	Breast	6585	3293	0.908 ± 0.006	0.540 ± 0.015a	0.278 ± 0.019	0.916 ± 0.005	0.916 ± 0.005
DC-AZ	Lung	2340	1170	0.526 ± 0.017	0.371 ± 0.011a	0.382 ± 0.031	0.589 ± 0.018	0.603 ± 0.016
DC-AZ	Urinary tract	2921	1461	0.744 ± 0.021	0.497 ± 0.014a	0.331 ± 0.025	0.798 ± 0.017	0.788 ± 0.018

^aLinear regression was used instead of logistic regression.