Table 2. Results of localization procedure with finite coverage.
| Distance | Rank | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| s | C | q .1 | q .25 | q .5 | q .75 | q .9 | q .1 | q .25 | q .5 | q .75 | q .9 | 𝔼(＃SS) |
| 0.02 | 10 | 5200 | 15880 | 33670 | 53400 | 67881 | 34 | 195 | 795 | 1496 | 1770 | 2085 |
| 0.02 | 30 | 3720 | 11460 | 28850 | 52730 | 77380 | 25 | 308 | 912 | 1499 | 1690 | 2084 |
| 0.02 | 5310 | 12419 | 28360 | 48630 | 60750 | 17 | 152 | 863 | 1420 | 1676 | 2082 | |
| 0.05 | 10 | 0 | 3170 | 18380 | 42240 | 63170 | 1 | 6 | 304 | 1161 | 1635 | 2080 |
| 0.05 | 30 | 0 | 0 | 14330 | 38860 | 57019 | 1 | 1 | 129 | 1356 | 1619 | 2086 |
| 0.05 | 990 | 9080 | 28750 | 53110 | 69750 | 2 | 4 | 159 | 1473 | 1639 | 2083 | |
| 0.10 | 10 | 0 | 0 | 3770 | 27300 | 55960 | 1 | 1 | 5 | 373 | 1591 | 2082 |
| 0.10 | 30 | 0 | 0 | 290 | 26950 | 50649 | 1 | 1 | 2 | 498 | 1539 | 2091 |
| 0.10 | 0 | 0 | 14079 | 37290 | 58970 | 1 | 1 | 3 | 493 | 1583 | 2082 | |
Data were generated as in Table 1 and then sampled to simulate sequencing. The number of homozygous founder lines was fixed to F = 200 in this study. Average coverage depth is indicated in the column labeled C. The rows denoted “” correspond to simulations in which each segregating site had a random level of coverage depth drawn from the empirical coverage distribution observed in actual E&R sequencing data. The column labeled q j corresponds to the jth percentile. The column labeled 𝔼(＃SS) shows the average number of segregating sites observed over all simulations. As the table shows, the additional noise introduced by low coverage depth makes the problem of localizing the selected site more challenging. However, under strong selection (s = 0.1), the true selected site was among the top five segregating sites in more than half the trials.