Analysis of family resemblance. 3. Complex segregation of quantitative traits.
Analysis of family resemblance. 3. Complex segregation of quantitative traits.
复制标题
家族相似性分析。
DOI：
--
复制
发表时间：
1974
影响因子：
9.8
通讯作者：
C. Maclean
中科院分区：
文献类型：
--
作者：
N. Morton;C. Maclean
在本系列的前几篇论文中，我们介绍了家庭相似性问题，包括分离和通径分析、杂交祖先家庭中表现的群体差异、连锁、突变筛查、亲子排斥和复发风险[1，2]。Wright的通径分析[3]和Fisher的归一化z变换[4]的结合被强调为区分遗传因素、共同和随机环境及其二阶效应(基因-环境相关性和分类交配)的一种手段，在交叉种族群体内和之间都是如此。通径分析提供了共同环境和遗传因素之间的最佳区分，而分离分析更好地区分了主基因座。在这里，我们将考虑在一个包含多基因、一个主基因座和随机环境和共同环境的模型下进行数量性状复杂分离分析。大多数分离分析的应用都涉及到一种二分法(正常和受影响的)，这不仅没有能力区分复杂的假设，而且也与遗传咨询的需要脱节[5]。一个担心糖尿病的人应该接受葡萄糖耐量测试，因为它比他亲属中糖尿病的发生具有更大的预测能力。此外，在家庭数据有用的情况下，亲属的数量特征往往比他们的情感状况更能提供信息。分析和咨询都应该清楚地使用定量和定性的信息。当完全量化不可行时，三分法(正常、中间、受影响)提供比两种状态更多的信息。除了主基因、多基因变异和环境效应外，我们的模型还包含了一个关于数量性状与情感关系的假设。埃尔斯顿和斯图尔特[6]讨论了混合模型，将其作为多基因和主基因情况的推广，并得出结论：“扩展到更一般的遗传模型不存在理论问题，但可能取决于计算机技术的实际进步。”我们发现，混合模式在当前的能力范围内。尽管不包括多个等位基因和两个或更多个主基因座这样的复杂情况，但该模型的优点是可管理的小
In previous papers of this series, we have introduced problems of family resemblance, which include segregation and path analysis, group differences expressed in families of hybrid ancestry, linkage, mutation screening, parentage exclusion, and recurrence risks [1, 2]. Combination of Wright's path analysis [3] with Fisher's normalizing z transformation [4] was stressed as a means to discriminate among genetic factors, common and random environment, and their second-order effects (gene-environment correlation and assortative mating), both within and among intercrossing racial groups. Path analysis provides optimal discrimination between common environment and genetic factors, while segregation analysis is better for distinguishing major loci. Here we shall consider complex segregation analysis of quantitative traits under a model which includes polygenes, a major locus, and both random and common environment. Most applications of segregation analysis have involved a dichotomy (normal versus affected), which not only has low power to discriminate complex hypotheses but is also out of touch with the needs of genetic counseling [5]. An individual apprehensive of diabetes should submit to a glucose-tolerance test because it has greater predictive power than the occurrence of diabetes among his relatives. Moreover, where familial data are useful, quantitative traits of relatives are often more informative than their affection status. Clearly quantitative as well as qualitative information should be used for both analysis and counseling. When full quantification is not feasible, a trichotomy (normal, intermediate, affected) provides more information than two states. In addition to a major locus, polygenic variation, and environmental effects, our model incorporates a hypothesis about the relation of the quantitative trait to affection. Elston and Stewart [6] discussed mixed models as a generalization of polygenic and major-locus cases and concluded that "extension to the more general genetic models presents no theoretical problems, but may depend on practical advances in computer technology." We have found that the mixed model is within current capabilities. Although not including such complications as multiple alleles and two or more major loci, this model has the advantage of a manageably small