Table 2.

Quadruplex-based ML models

	Quadruplex-based models
Cancer type	Median test AUC	Percentage of datasets with the mean test AUC confidence interval not containing 0.5	Median lift of recall	Percentage of datasets with lift of recall higher than 1.5
skin	0.58	100	1.69	70
overall	0.55	100	1.58	61
prostate	0.54	100	1.00	9
uterus	0.58	100	1.76	79
bone	0.70	100	3.17	100
brain	0.56	100	1.63	65
breast	0.65	100	3.64	100
ovary	0.56	100	1.31	45
pancreatic	0.54	100	1.21	9
blood	0.55	100	1.32	40
liver	0.55	100	1.13	57

Performance metrics by cancer type