Table 1.

Characteristics of data sets.

Data set	Ref	Data type	# features	Training set		Independent validation set
				# cases	# control
Leukemia	13	Gene expression	3,051	11a	27	No
Lung cancer	19	Gene expression	12,533	16	16	149 (15 controls, 134 cases)
Prostate cancer	20	Proteomics	15,154	30	30	262 (223 controls, 39 cases

Note:

^aPatients with acute myeloid leukemia were considered “cases” and those with acute lymphoblastic leukemia were used as “controls”.