Type I error rates on the ENCODE3 datasets.

Notes: Seven ENCODE3 regions are available for analysis. In each region, we tested the overrepresentation of rare variants between the CHB (90 individuals) and the YRI populations (120 individuals). The MAF cutoff 1% was applied. For each dataset, the total number of variants was given. To evaluate the type I error rate, we permuted each individual's population attribute and analyzed the permuted datasets. Type I error rates were estimated on 1,000 replicates at the significance level 0.05. See notes for table 1 for abbreviation detail.