Detecting purely epistatic multi-locus interactions by an omnibus permutation test on ensembles of two-locus analyses

Table 12 Summary of prediction accuracy by MDR from early genetic association studies of T2D in a Korean population, a Han Chinese population from Taiwan, a female population from the US, and that from an early genetic association study of metabolic syndrome in an Italian population from the Centre East Coast Italy.

Reference	Population	Gene	Prediction Accuracy	CVC	Permutation p-value
Cho et al. [43]	Korean	PPARG, UCP2	0.7957	9/10	0.01
Hsieh et al. [44]	Han Chinese	RXRG, EGFR	0.6270	11/12	N/A
Qi et al. [45]	US	KCNJ11, HNF4A	0.5420	10/10	0.010
Fiorito et al. [47]	Italian	PPARG, DIO2	0.6170	10/10	0.005

A permutation test with 1,000 randomised replicates is performed to obtain the empirical p-value for the null hypothesis of no association in the studies conducted in the US and Italian populations. In contrast, a permutation test with 100 randomised replicates is performed to obtain the empirical p-value in the study conducted in the Korean population.

ISSN: 1471-2105