BMC Bioinformatics

Table 1 Mean percentage of significant cases detected when the null hypothesis was always true.

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

	N	S	S _α	SB	BH	SGoF
t-test	5	100	4.95 ± 0.421	0.042 ± 0.0139	0.046 ± 0.0171	0.047 ± 0.0306
	5	1000	5.02 ± 0.237	0.006 ± 0.0016	0.007 ± 0.0020	0.008 ± 0.0265
	5	10000	5.00 ± 0.129	0.001 ± 0.0002	0.001 ± 0.0002	0.002 ± 0.0116
	10	100	4.97 ± 0.412	0.063 ± 0.0184	0.068 ± 0.0209	0.042 ± 0.0485
	10	1000	4.97 ± 0.236	0.005 ± 0.0015	0.005 ± 0.0016	0.007 ± 0.0275
	10	10000	5.00 ± 0.130	0.000 ± 0.0001	0.000 ± 0.0002	0.002 ± 0.0079
	20	100	5.03 ± 0.419	0.045 ± 0.0157	0.052 ± 0.0194	0.045 ± 0.0294
	20	1000	4.99 ± 0.233	0.006 ± 0.0021	0.006 ± 0.0025	0.004 ± 0.0076
	20	10000	5.02 ± 0.130	0.001 ± 0.0001	0.001 ± 0.0002	0.002 ± 0.0087
			Mean	0.019 ± 0.0242	0.021 ± 0.0267	0.018 ± 0.0204
	20	100	5.16 ± 0.421	0.049 ± 0.0149	0.052 ± 0.0166	0.046 ± 0.0356
	20	1000	5.21 ± 0.239	0.004 ± 0.0014	0.004 ± 0.0017	0.014 ± 0.0157
G-test	20	10000	5.22 ± 0.133	0.001 ± 0.0001	0.001 ± 0.0003	0.019 ± 0.0558
	40	100	5.05 ± 0.427	0.084 ± 0.0217	0.092 ± 0.0282	0.066 ± 0.0546
	40	1000	5.12 ± 0.239	0.003 ± 0.0014	0.004 ± 0.0021	0.008 ± 0.0108
	40	10000	5.09 ± 0.132	0.001 ± 0.0002	0.001 ± 0.0002	0.005 ± 0.0215
			Mean	0.024 ± 0.0350	0.026 ± 0.0381	0.026 ± 0.0243

Values are averages through 1,000 replicates and their ± standard deviations. The simulated null models were tested under t or homogeneity G tests (see text). N: sample size. S: number of tests. S_α: Percentage of significant tests. SB: Sequential Bonferroni. BH: Benjamini and Hochberg. SGoF: Sequential Goodness of Fit.

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com