Figure 7: The mean computation time (in seconds) of every method, by varying the region length, the sample size, or the number of permutations (for ADA) or random subsamples (for BE-SKAT and BE-BURDEN).

Left column: sample size =1000, the number of permutations or random subsamples =1000; middle column: region length =5 kb, the number of permutations or random subsamples =1000; right column: region length =5 kb, sample size =1000.