Confident difference criterion: a new Bayesian differentially expressed gene selection algorithm with applications

Table 1 Performance evaluation under Study I (Setting 1), (G=5000, 500 DE gene) ^#

Cut-off	Method	CDE	CCDE	CCEE	FNR	FPR	FDR	FNDR
γ ₀	I	796.7(14.5)	466.7(4.7)	4169.9(13.9)	0.067(0.009)	0.073(0.003)	0.414(0.011)	0.008(0.001)
(0.4)	II	864.4(13.8)	499.2(1.1)	4134.8(13.8)	0.002(0.002)	0.081(0.003)	0.422(0.009)	0.000(0.000)
γ ₀	I	526.5(9.8)	419.2(6.7)	4392.7(6.9)	0.162(0.013)	0.024(0.002)	0.204(0.011)	0.018(0.001)
(0.6)	II	582.3(7.3)	493.3(3.0)	4411.0(6.9)	0.013(0.006)	0.020(0.002)	0.153(0.010)	0.002(0.001)
FDR	I	296.4(13.7)	283.3(12.9)	4486.9(2.9)	0.433(0.026)	0.003(0.001)	0.044(0.009)	0.046(0.003)
(0.05)	II	469.2(13.1)	450.7(10.2)	4481.5(4.4)	0.099(0.020)	0.004(0.001)	0.039(0.009)	0.011(0.002)
	SAM	330.5(16.0)	314.0(13.5)	4483.4(4.7)	0.372(0.027)	0.004(0.001)	0.050(0.013)	0.040(0.003)
	LIMMA	320.2(15.2)	304.9(13.7)	4484.7(4.1)	0.390(0.027)	0.003(0.001)	0.048(0.012)	0.042(0.003)
	SPH	192.0(10.6)	188.1(10.0)	4496.1(2.1)	0.624(0.020)	0.001(0.000)	0.020(0.011)	0.065(0.002)
	EBA	166.4(14.1)	158.8(13.3)	4492.3(2.2)	0.682(0.027)	0.002(0.000)	0.046(0.012)	0.071(0.003)

^#Empirical estimates of the standard deviation were reported in the parentheses

ISSN: 1471-2105