Average properties (with standard errors) of the 50 data sets in the simulation study underlying Tables 2, 3, 4, 5, 6
FST/struct | FST (%) | |#cont – #case| | λ |
|---|---|---|---|
| (1) No structure | — | — | 1.2 (0.03) |
| (2) Low/eq | 3.5 (0.1) | 24 (2) | 1.2 (0.04) |
| (3) High/eq | 8.3 (0.1) | 43 (4) | 1.9 (0.19) |
| (4) Low/un | 3.5 (0.1) | 19 (2) | 1.2 (0.05) |
| (5) High/un | 8.4 (0.1) | 20 (2) | 1.4 (0.06) |
[i] The first column gives a short label for each of the five demographic scenarios described in the text. Columns 2–4 give the average maximum-likelihood estimate of FST; the average absolute difference between the numbers of cases and of controls drawn from one population; and the average estimate of the λ parameter of the genomic control method.