Finally, Figure S5 compares the distribution of CSS and its component tests (FST, ΔDAF, and XP-EHH) at the 12 candidate regions harboring stature-associated genes. These comparisons show that the CSS scores are not being derived from one particular component selection test. The results also suggest that the three component tests provide complementary evidence to each other at the candidate regions. However, it is evident that many of these candidate regions would not have been detected with any of the component selection tests, particularly regions 1, 2, 5, 8, 9, and 12. These results provide overwhelming support for the CSS strategy, suggesting that it is a robust and efficient method for combining complementary evidence to detect the historical signatures of selection for complex traits, which otherwise cannot be detected by available tests of selection.