. 2018 Sep 10;19:317. doi: 10.1186/s12859-018-2318-8

Table 3.

Simulation study 1, scenario 1 with n_X=30, n_Y=10 and 100 runs

			Δ=1.5
Nb. of			α
DE genes		0.1	0.05	0.01
50	${\hat{p}}_{m}$	0.85	1	1
	%	97.2 (2.8)	91.1 (2.1)	69.6 (6.7)
100	${\hat{p}}_{m}$	1	1	1
	%	93.7 (2.5)	86.0 (3.5)	60.1 (5.0)
200	${\hat{p}}_{m}$	1	1	1
	%	84.9 (2.6)	70.5 (3.5)	34.9 (3.4)
			Δ=2
Nb. of			α
DE genes		0.1	0.05	0.01
50	${\hat{p}}_{m}$	0.13	0.33	0.98
	%	99.7 (0.7)	99.2 (1.2)	93.2 (3.6)
100	${\hat{p}}_{m}$	0.38	0.88	1
	%	99.6 (0.6)	98.4 (1.0)	88.4 (3.2)
200	${\hat{p}}_{m}$	0.99	1	1
	%	97.8 (1.1)	92.1 (1.9)	65.3 (4.0)
			Δ=3
Nb. of			α
DE genes		0.1	0.05	0.01
50	${\hat{p}}_{m}$	0.00	0.00	0.02
	%	100	100	96.0 (0.3)
100	${\hat{p}}_{m}$	0.00	0.00	0.14
	%	100	100	99.8 (0.5)
200	${\hat{p}}_{m}$	0.01	0.11	0.97
	%	100.0 (0.1)	99.9 (0.2)	98.1 (1.1)

Evaluation of the first step of the ORdensity method using different values of α. The Table shows the estimated probability, ${\hat{p}}_{m}$ , of no considering as a potential DE gene at least one gene that it really is, and the mean proportion of DE genes (row named “%”) that the procedure considered as potential DE genes. Corresponding standard deviations are in brackets