A novel approach to minimize false discovery rate in genome-wide data analysis

BMC Systems Biology

Table 1 Null and alternative hypotheses in simulated datasets

Category	# of features	Group 1	Group 2
1	5000	Gaussian with mean = 0 and variance = 1	Gaussian with mean = 0 and variance = 1
2	5000	Uniform in range $[- \sqrt{3,} \sqrt{3}]$	Uniform in range $[- \sqrt{3,} \sqrt{3}]$
3	50	Gaussian with mean = 0 and variance = 1	Gaussian with mean = -2 and variance = 1
4	150	Gaussian with mean = 0 and variance = 1	Gaussian with mean = 1 and variance = 1
5	150	Uniform in range $[- \sqrt{3,} \sqrt{3}]$	Uniform in range $[1 - \sqrt{3,} 1 + \sqrt{3}]$
6	50	Uniform in range $[- \sqrt{3,} \sqrt{3}]$	Uniform in range $[1.5 - \sqrt{3,} 1.5 + \sqrt{3}]$

Each simulation dataset has 10400 features, which are divided into six groups. The number of features in each group and the corresponding distributions are listed below. Category 1 and 2 are for null hypotheses features and category 3, 4, 5 and 6 are for alternative features. Particularly in Category 2, we use uniform distribution in range $[- \sqrt{3,} \sqrt{3}]$ to make the variance equal to 1.

ISSN: 1752-0509