Permutation – based statistical tests for multiple hypotheses

Table 2 Results of analyses of statistical tests.

Test	Data description	Groups	Features	Samples	Feature selection according to
					raw P-value	Multiple test correction
						Bonferroni	B&H	PT
Bartlett	Microarray Numerical	2	8068	12	526	1	1	327
t-test	Microarray Numerical	2	8068	12	1413	2	39	1398
Chi-square	Single nucleotide polymorphisms (SNP) Categorical	3	334	33	153	8	131	153
ANOVA	Microarray Numerical	3	14976	37	6371	9	3331	6262

Comparative statistics when multiple-testing correction is based on either Bonferroni correction, Benjamini and Hochberg (B&H), or permutation test (PT), against potentially significant features (P-value) before correction. Significance level (α) = 0.05, Number of permutations (B) = 10000.

ISSN: 1751-0473