Applications of Different Weighting Schemes to Improve Pathway-Based Analysis

Table 4

Global test results on simulated datasets under RWM scheme and uniform weight: (a) simulation case 1 uses covariance structure and mean of Pathway ID 164 from type II diabetes dataset, (b) simulation case 2 uses covariance structure and mean of Pathway ID 441 from canine dataset.

(a) Simulation Case 1.

No. of samples	No. of genes	No. of tested	Statistic Q		Expected Q	sd-of-Q		P values
			Unif	RWM		Unif	RWM	Unif	RWM
30	26	26	13.51	27.68	10	7.64	3.96	0.2246	0.0014
50	26	26	13.42	27.49	10	6.59	3.61	0.2155	0.0007
100	26	26	12.33	27.44	10	6.60	3.25	0.2573	0.0002

(b) Simulation Case 2.

No. of samples	No. of genes	No. of tested	Statistic Q		Expected Q	sd-of-Q		P values
			Unif	RWM		Unif	RWM	Unif	RWM
30	21	21	34.60	46.51	10	9.46	5.57	0.0289	0.0002
50	21	21	66.41	56.91	10	7.94	5.87	0.0004	0.0001
100	21	21	97.11	70.69	10	8.03	5.42	0.0000	0.0000