Marginal Count

n_.j

rr

Power

Sensitivity

FDR

Scenario 1: Select one signal in one drug with marginal total as 28,216 or 4,362

28216

500

3

0.073

0.13

0.0565

28216

1500

3

0.315

0.278

0.0442

28216

3000

3

0.83

0.753

0.044

28216

5000

3

0.992

0.983

0.0368

28216

8000

3

1

1

0.0257

28216

12000

1

0.061

NA

0.053

28216

12000

2

0.758

0.751

0.043

28216

12000

3

1

1

0.0351

28216

12000

5

1

1

0.0326

28216

12000

7

1

1

0.0316

4362

12000

1

0.077

NA

0.077

4362

12000

2

0.084

0.005

0.090

4362

12000

3

0.152

0.079

0.073

4362

12000

4

0.50

0.364

0.0569

4362

12000

5

1

1

0.0393

4362

12000

7

1

1

0.0397

Scenario 2: Randomly select 30 common signals with homogenerous rr

NA

AS

1

0.062

NA

0.064

NA

AS

2

0.744

0.02107

0.0381

NA

AS

3

1

0.1347

0.0057

NA

AS

5

1

0.4794

0.0014

NA

AS

7

1

0.7315

0.0011

NA

AS

9

1

0.8520

0.0009

Scenario 3: Randomly select 30 signals (not necessarily common) with homogenerous rr

NA

AS

1

0.085

NA

0.081

NA

AS

2

0.80

0.0163

0.0293

NA

AS

3

1

0.1426

0.0050

NA

AS

5

1

0.4450

0.0018

NA

AS

7

1

0.6829

0.0013

NA

AS

9

1

0.8018

0.0010

Scenario 4: Randomly select 30 signals with inhomogenerous rr

NA

AS

1

0.056

NA

0

NA

AS

3

1

0.93912

0.0009

NA

AS

5

1

0.98467

0.006

NA

AS

7

1

0.99887

0.018