Table 1. Detailed overview of the different scenarios considered for the type I error and power study.

Investigation	Scenario	% of causal variants	Effect size weights	Direction of effect: Positive / Negative	Median (MAD) of explained variance in %
Type I error	0	0%	c = 0	-	-
Power	1	5%	c = 0.6	100% / 0%	0.9% (0.6)
	2	5%	c = 0.3	100% / 0%	0.2% (0.2)
	3	5%	c = 0.2	100% / 0%	0.1% (0.1)
	4	10%	c = 0.6	100% / 0%	1.9% (1.4)
	5	10%	c = 0.3	100% / 0%	0.5% (0.3)
	6	10%	c = 0.2	100% / 0%	0.2% (0.2)
	7	20%	c = 0.6	100% / 0%	3.8% (2.1)
	8	20%	c = 0.3	100% / 0%	1.0% (0.6)
	9	20%	c = 0.2	100% / 0%	0.4% (0.2)
	10	50%	c = 0.6	100% / 0%	9.1% (3.0)
	11	50%	c = 0.3	100% / 0%	2.4% (0.9)
	12	50%	c = 0.2	100% / 0%	1.1% (0.4)
	13–24	As in scenarios 1–12	As in scenarios 1–12	80% / 20%	As in scenarios 1–12
	25–36	As in scenarios 1–12	As in scenarios 1–12	50% / 50%	As in scenarios 1–12

The scenarios vary the percentage of causal variants, their effect size, and the percentage of causal variants with effects in positive/ negative direction. Scenarios 13–24 and 25–36 have the same percentage of causal SNVs and the same effect sizes as scenarios 1–12, but 80% / 20% and 50% / 50% of effects in positive / negative direction. The percentage of causal rare variants is with respect to the total number of rare variants with MAF≤0.03 in the gene. The effect size of a variant with a given MAF on the trait Y is β_i = c ∙ |log₁₀(MAF_i)|. The percentage of explained variance for a given gene is calculated as the sum of $2\bullet {MAF}_i\bullet \left(1-{MAF}_i\right)\bullet {\beta }_i^2/Var\left(Y\right)$ over all variants i in the gene. Reported are the median and the median absolute deviation (MAD) of this heritability estimate over the 10,000 replicates.