Table 1.

Type I Error Estimates of SKAT Aimed at Testing an Association between Randomly Selected 30 kb Regions with a Continuous Trait at Type I Error Rates as Low as the Genome-wide α = 10⁻⁶ Level

Total Sample Size (n)	Continuous Phenotypes			Dichotomous Phenotypes
	α = 10−4	α = 10−5	α = 10−6	α = 10−4	α = 10−5	α = 10−6
500	7.4 × 10−5	6.5 × 10−6	5.9 × 10−7	2.2 × 10−5	1.0 × 10−6	1.0 × 10−8
1000	8.5 × 10−5	8.2 × 10−6	8.0 × 10−7	5.0 × 10−5	3.5 × 10−6	2.3 × 10−7
2500	9.6 × 10−5	9.1 × 10−6	8.4 × 10−7	7.6 × 10−5	6.3 × 10−6	5.6 × 10−7
5000	9.8 × 10−5	9.6 × 10−6	8.8 × 10−7	8.9 × 10−5	7.8 × 10−6	7.0 × 10−7

Each entry represents type I error rate estimates as the proportion of p values α under the null hypothesis based on 10⁸ simulated phenotypes.