Discovery of dominant and dormant genes from expression data using a novel generalization of SNR for multi-class problems

Table 2 Details of the selected genes by the frequency-based method for the Leukemia data set

	Class	Probe ID	Gene Symbol	Frequency	Ave. GDI	p-value	q-value
Dom	ALL	1389_at	MME	96	1.98	0	0
		32847_at	MYLK	62	2.02	0	0
		32872_at	ESTs	58	1.73	0	0
		35164_at	WFS1	52	2.05	0	0
		37280_at	SMAD1	25	1.87	0	0
	MLL	34306_at	MBNL1	99	1.43	0	0
		40763_at	MEIS1	92	1.41	0	0
		36777_at	KLRK1	83	1.42	0	0
		1065_at	FLT3	56	1.20	0	0
		34583_at	FLT3	30	1.30	0	0
	AML	39566_at	CHRFAM7A	46	1.89	0	0
		41752_at	GHITM	39	1.51	0	0
		38710_at	OTUB1	31	1.54	0	0
		37187_at	CXCL2	22	1.53	0	0
		36162_at	BSG	21	1.47	0	0
Dor	ALL	33412_at	LGALS1	94	1.66	0	0
		37403_at	ANXA1	90	1.74	0	0
		37809_at	HOXA9	62	1.59	0	0
		41448_at	HOXA10	54	1.64	0	0
		31575_f_at	ESTs	33	1.65	0	0
	MLL	1674_at	YES1	69	1.03	0	0
		1325_at	SMAD1	43	0.97	9.54E-07	2.80E-05
		539_at	RYK	39	1.05	0	0
		1971_g_at	FHIT	33	0.90	9.54E-07	3.02E-05
		37527_at	ELK3	28	0.98	9.54E-07	2.81E-05
	AML	41747_s_at	MEF2A	50	1.89	0	0
		41503_at	ZHX2	49	1.88	0	0
		37988_at	CD79B	37	1.98	0	0
		37710_at	MEF2C	34	2.04	0	0
		40966_at	STK39	32	2.12	0	0

Using the training data, for each class the top five most frequently occurring genes are selected.

ISSN: 1471-2105