Table 3.
Expected risk scores based on the naïve and genome‐wide (GW) bootstrap mean logHR estimates from 5,000 simulation datasets compared to the risk score based on the generating values
| Expected risk score | |||||
|---|---|---|---|---|---|
| Simulation study | Bootstrap P‐value threshold | Generating | Naïve | GW | Naïve – GW difference |
| 1 | 5 × 10−5 | 3.49 | 4.44 | 3.40 | 1.04 |
| 2 | 0.01 | 3.49 | 4.18 | 2.83 | 1.35 |
| 3 | 5 × 10−5 | 2.89 | 4.38 | 3.45 | 0.93 |
Risk scores are calculated from 15 SNPs generated to have association with risk of severe nephropathy in a sample of 5,444 individuals. Expected scores are taken over a hypothetical population with the same marginal allele frequencies as the sample of 5,444.