We acknowledge that smoother measures than thresholding are available. For example, the Rank-Rank-Hypergeometric-Overlap (RRHO) test42 displays overlap in a more nuanced manner. However, as we generally worked with 13 phenotypes, resulting in 13 × 13–13 = 156 combinations (considering that some of the thresholded tests were not symmetrical), a method like RRHO would yield 156 individual figures from which it would be difficult to discern patterns in the results.
