Thompson E A
Am J Hum Genet. 1981 Nov;33(6):968-77.
Criteria and rules are derived for the most statistically efficient sampling of relatives of affected probands under a simple class of standard genetic models. Results show that the optimal relative type depends on the alternative hypotheses of particular interest. Thus, there is no uniformly optimal strategy, but the criteria can be used to derive the best strategy in any particular case; examples are given.
在一类简单的标准遗传模型下,得出了对患病先证者亲属进行统计学上最有效抽样的标准和规则。结果表明,最佳亲属类型取决于特定感兴趣的备择假设。因此,不存在统一的最优策略,但这些标准可用于推导任何特定情况下的最佳策略;文中给出了示例。