Table 3.
Sample size required to detect genetic associations combining different family structures for high penetrances*
| ε = 0 | ε = 0.005 | ε = 0.01 | |
|---|---|---|---|
| Dominant | |||
| p = 0.05 | 338(373) | 537(500) | 2124(∞**) |
| p = 0.20 | 185(217) | 202(232) | 269(290) |
| p = 0.70 | 1746(2104) | 4663(5890) | ∞ (∞) |
| Recessive | |||
| p = 0.05 | 48005(45598) | ∞ (∞) | ∞ (∞) |
| p = 0.20 | 1139(1126) | 2238(2121) | ∞ (∞) |
| p = 0.70 | 155(159) | 167(170) | 215(216) |
| Multiplic. | |||
| p = 0.05 | 1513(1669) | ∞ (∞) | ∞ (∞) |
| p = 0.20 | 443(486) | 564(587) | 1458(1568) |
| p = 0.70 | 333(358) | 385(409) | 704(724) |
| Additive | |||
| p = 0.05 | 869(967) | 4448(9965) | ∞ (∞) |
| p = 0.20 | 335(376) | 398(430) | 762(752) |
| p = 0.70 | 485(534) | 598(650) | 1804(1867) |
*
The values in brackets are based on the weighting scheme suggested by Risch & Teng (1998);
**
∞ means that 80% power cannot be attained or the sample size required is unrealistically large (greater than 100 000); Significance level α = 5 × 10−8; power 1 − β = 0.80; Dominant model: f2 = f1 = 0.4, f0 = 0.1; Recessive model: f2 = 0.4, f1 = f0 = 0.1; Multiplicative model: f2 = 0.4, f1 = 0.2, f0 = 0.1; Additive model: f2 = 0.4, f1 = 0.25, f0 = 0.1.