A new multitest correction (SGoF) that increases its statistical power when increasing the number of tests

BMC Bioinformatics

Table 2 Percentages of significant cases detected (Detected) and false discovery rate (FDR) after multitest adjustment when the p-values come from families of one-sample t tests where some (% effect) of the alternative hypotheses were true.

	Weak			SB		BH		SGoF
N	% effect	S	Significant	Detected	FDR	Detected	FDR	Detected	FDR
5	5%	100	5.2	0.04	98	0.04	98	0.07	85
5	5%	1000	5.2	0.01	78	0.01	80	0.01	91
5	5%	10000	5.2	0.00	91	0.00	88	0.02	89
5	10%	100	5.4	0.06	86	0.06	87	0.09	83
5	10%	1000	5.5	0.01	83	0.01	84	0.04	78
5	10%	10000	5.5	0.00	80	0.00	83	0.11	80
5	20%	100	5.9	0.08	57	0.08	60	0.12	64
5	20%	1000	6.0	0.01	57	0.01	58	0.13	63
5	20%	10000	6.0	0.00	70	0.00	69	0.54	63
			Mean	0.02	77.8	0.02	78.6	0.12	77.3
			SD	0.030	14.18	0.030	13.45	0.161	11.24
10	5%	100	5.7	0.05	72	0.06	70	0.09	80
10	5%	1000	5.6	0.01	69	0.01	67	0.06	77
10	5%	10000	5.6	0.00	71	0.00	71	0.22	74
10	10%	100	6.2	0.06	56	0.08	57	0.19	67
10	10%	1000	6.2	0.01	44	0.01	47	0.24	59
10	10%	10000	6.3	0.00	44	0.00	45	0.84	61
10	20%	100	7.6	0.12	35	0.15	35	0.52	42
10	20%	1000	7.5	0.01	30	0.02	30	1.14	42
10	20%	10000	7.5	0.00	27	0.00	30	2.11	44
			Mean	0.03	49.8	0.04	50.2	0.60	60.7
			SD	0.041	17.85	0.051	16.72	0.674	15.15
20	5%	100	6.4	0.14	33	0.17	35	0.20	60
20	5%	1000	6.5	0.02	31	0.03	30	0.33	51
20	5%	10000	6.4	0.00	18	0.00	19	1.01	53
20	10%	100	7.8	0.24	17	0.32	18	0.61	38
20	10%	1000	7.9	0.03	13	0.06	16	1.53	36
20	10%	10000	7.9	0.01	7	0.01	10	2.46	40
20	20%	100	10.7	0.36	8	0.57	10	2.24	24
20	20%	1000	10.8	0.06	6	0.15	9	4.38	25
20	20%	10000	10.8	0.01	4	0.04	6	5.35	27
			Mean	0.10	15.22	0.15	17	2.01	39.3
			SD	0.126	10.67	0.188	9.89	1.81	13.02

Weak: The alternative hypothesis is weak implying that data come from a Normal(0.36, 1). N: sample size. % effect: % of real true alternatives. S: number of tests. Significant: % of significants before adjustment. Detected: % of significant tests after adjustment. FDR: false discovery rate. SB: Sequential Bonferroni. BH: Benjamini and Hochberg. SGoF: Sequential Goodness of Fit.

ISSN: 1471-2105