Table 3.

Robustness of the proposed correction

True correlation	Plug-in correlation	TP	FP	FDP
0.4	0.3	261.16 (260.27, 262.85)	2.42 (2.12, 2.71)	0.0091 (0.0080, 0.0102)
0.4	0.35	252.20 (250.92, 253.48)	1.56 (1.32, 1.80)	0.0061 (0.0052, 0.0070)
0.4	0.375	247.78 (246.79, 249.06)	1.48 (1.22, 1.73)	0.0059 (0.0049, 0.0069)
0.4	0.4	243.59 (242.31, 244.879)	1.40 (1.17, 1.63)	0.0057 (0.0048, 0.0066)
0.4	0.425	238.72 (237.42, 240.02)	1.60 (1.38, 1.82)	0.0066 (0.0057, 0.0075)
0.4	0.45	235.11 (233.88, 236.34)	1.96 (1.72, 2.20)	0.0082 (0.0072, 0.0092)
0.4	0.5	234.81 (233.57, 236.04)	1.96 (1.72, 2.20)	0.0082 (0.0072, 0.0092)

For the “positive pleiotropy A” scenario the correlation due to overlap is 0.4. Here we varied the correlation value in the de-correlation step from 0.3 to 0.5. TP, true positives; FP false positives, FDP, false discovery proportion