Table 1 Effect of sample sizes on the results of epistatic QTL analysis (100 replicates)

From: A penalized maximum likelihood method for estimating epistatic effects of QTL

 

QTL

Main effect

Interaction

σ2

  

Power

Position

Effect

Power

Position 1

Position 2

Effect

 

1

True value

0.00

1.0000

0.00

80.00

1.0000

10.0000

 

n=200

63

3.02(5.86)

1.0507(0.2852)

51

4.86(9.61)

82.38(25.81)

0.9975(0.2506)

10.7410(1.4764)

 

n=500

92

1.22(4.41)

0.9104(0.1859)

93

0.72(3.05)

78.44(8.77)

0.9427(0.1879)

9.9434(0.7291)

2

True value

60.00

1.4142

40.00

140.00

1.4142

 
 

n=200

85

58.71(5.30)

1.4993(0.2908)

87

42.17(13.50)

137.74(14.33)

1.4178(0.3755)

 
 

n=500

100

60.23(1.79)

1.4247(0.1804)

100

39.24(4.78)

139.50(5.14)

1.3787(0.1856)

 

3

True value

140.00

2.0000

80.00

200.00

2.0000

 
 

n=200

100

139.75(2.65)

1.9811(0.2906)

98

79.53(3.65)

199.37(2.71)

1.9659(0.3050)

 
 

n=500

100

140.01(0.09)

1.9865(0.1634)

100

79.98(0.27)

200.00(0.00)

2.0079(0.1684)

 

4

True value

180.00

2.8284

100.00

120.00

2.8284

 
 

n=200

100

179.89(1.86)

2.8074(0.3012)

98

99.30(3.61)

120.10(1.67)

2.6130(0.4750)

 
 

n=500

100

179.99(0.08)

2.8205(0.1856)

100

99.98(0.16)

120.00(0.00)

2.8166(0.2048)

 
  1. n=200 and n=500 represent estimates from sample sizes 200 and 500, respectively.
  2. The standard deviations of estimates are calculated from only the significant samples and are given in parentheses.