Department of Genetics, University of Georgia, Athens, Georgia 30602.
Genetics. 1985 Apr;109(4):785-98. doi: 10.1093/genetics/109.4.785.
In population studies, adults are frequently difficult or inconvenient to identify for genotype, but a family profile of genotypes can be obtained from an unidentified female crossed with a single unidentified male. The problem is to estimate an allele frequency in the cryptic parental gene pool from the observed family profiles. For example, a worker may wish to estimate inversion frequencies in Drosophila; inversion karyotypes are cryptic in adults but visible in salivary gland squashes from larvae. A simple mixture model, which assumes the Hardy-Weinberg law, Mendelian laws and a single randomly chosen mate per female, provides the vehicle for studying three competing estimators of an allele frequency. A simple, heuristically appealing estimator called the Dobzhansky estimator is compared with the maximum likelihood estimator and a close relative called the grouped profiles estimator. The Dobzhansky estimator is computationally simple, consistent and highly efficient and is recommended in practice over its competitors.
在群体研究中，成年人的基因型往往难以或不便确定，但可以从一个与单个未识别的雄性杂交的未知女性那里获得基因型的家族谱。问题是要从观察到的家族谱中估计隐性亲本基因库中的等位基因频率。例如，研究人员可能希望估计果蝇中的倒位频率；在成体中，倒位染色体组型是隐性的，但在幼虫的唾液腺压片中是可见的。一个简单的混合模型，假设哈迪-温伯格定律、孟德尔定律和每个雌性随机选择一个配偶，为研究三种竞争的等位基因频率估计器提供了工具。一个简单的、直观上吸引人的估计器称为多布赞斯基估计器，它与最大似然估计器和一个称为分组谱估计器的近亲进行了比较。多布赞斯基估计器计算简单、一致且高效，在实践中优于其竞争对手。
