Table 1.

Result of our approach on simulated dataset

Conditions	# of genes	# of samples	Direction	Se priority		Balanced		PPV priority
				PPV	Se	PPV	Se	PPV	Se
	10	100	u	0.85	0.92	0.87	0.87	0.91	0.26
Global			d	0.41	0.89	0.51	0.78	0.71	0.63
	100	1000	u	0.20	0.95	0.30	0.28	0.99	0.05
			d	0.11	0.93	0.25	0.16	0.93	0.09
	10	10	u	-	-	0.58	0.85	-	-
Local			d	0.57	0.93	0.78	0.78	0.94	0.63
	100	100	u	0.25	0.92	0.57	0.50	0.95	0.10
			d	0.18	0.89	0.33	0.60	0.97	0.10
	10	100	u	0.47	0.97	0.67	0.79	0.93	0.63
Dynamic			d	0.49	0.96	0.64	0.67	0.92	0.48
	100	1000	u	0.21	0.90	0.42	0.23	0.55	0.10
			d	0.11	0.91	0.39	0.21	0.94	0.11

Both dynamic and steady-state (Global and local) expression dataset has been tested. Please refer to Section “Comparison with other methods” for the “conditions” description of “Global”, “Local” and “Dynamic”. PPV: positive predicted value; Se: sensitivity; u: undirected; d: directed.