Skip to main content

Table 2 Comparison results among RareProb-C and the other 5 state-of-the-art approaches on the type-II error rates

From: An improved burden-test pipeline for identifying associations from rare germline and somatic variants

PAR

Causal

Type-II error

Approach

RareProb-C

LRT

LBL

BUR 0.95

LiMB

CCRS

0.02

 

12.89%

100%

32.99%

23.93%

42.78%

53.17%

0.03

50

15.21%

100%

30.09%

14.96%

57.37%

54.00%

0.04

 

17.28%

100%

19.78%

11.16%

47.99%

52.02%

0.05

 

20.90%

100%

25.51%

12.03%

52.76%

52.42%

0.02

 

12.20%

100%

26.06%

20.15%

56.10%

54.45%

0.03

60

17.05%

100%

26.18%

16.16%

60.48%

53.03%

0.04

 

20.75%

100%

39.02%

14.87%

65.71%

54.08%

0.05

 

23.29%

100%

31.14%

19.44%

58.82%

55.45%

0.02

 

12.29%

100%

31.88%

20.33%

67.80%

52.53%

0.03

70

19.10%

100%

33.11%

20.66%

67.94%

53.88%

0.04

 

22.39%

100%

33.32%

23.55%

71.77%

55.62%

0.05

 

22.51%

100%

32.09%

31.37%

75.55%

58.84%

0.02

 

11.39%

100%

32.66%

9.29%

80.07%

50.16%

0.03

80

17.85%

100%

48.38%

8.38%

76.71%

53.03%

0.04

 

22.98%

100%

38.21%

31.35%

84.29%

62.57%

0.05

 

15.57%

100%

42.83%

39.47%

78.66%

67.21%

0.02

 

11.79%

100%

36.91%

35.63%

91.26%

51.73%

0.03

90

20.42%

100%

40.49%

40.33%

89.29%

61.25%

0.04

 

21.28%

100%

39.91%

47.99%

91.26%

70.83%

0.05

 

14.72%

100%

48.87%

55.54%

88.79%

74.26%

  1. The population attributed risk (PAR) still vaires from 0.02 to 0.05 and the number of preset causal variants enumerates from 50 to 90. The significance threshold is set to P<0.05