Moldin S O, Rice J P, Van Eerdewegh P, Gottesman I I, Erlenmeyer-Kimling L
Department of Psychiatry, Washington University School of Medicine, St. Louis, MO.
Genet Epidemiol. 1990;7(5):371-86. doi: 10.1002/gepi.1370070507.
Adjunct consideration of both qualitative (affection status) and quantitative (correlated liability indicator) information to define a bivariate phenotype can increase considerably the accuracy and efficiency of disease risk estimation. A general approach for calculating morbid risks to offspring on the basis of parental affection status and an offspring quantitative trait is presented. We also describe two different bivariate models of multifactorial inheritance, as implemented in the computer programs POINTER and YPOINT, and make explicit their assumptions/constraints when estimating the within-person and parent-offspring correlations necessary for calculation of morbid risks. We use psychometric family data on schizophrenia from the New York High-Risk Project to estimate these correlations and illustrate our methods. Our results show that even when a trait is only moderately correlated with liability, incorporation of quantitative trait information can lead to resolution of a range of risk to offspring that is not possible through reliance on parental affection status alone. Bivariate models provide a useful methodology for incorporating quantitative indicators of liability in the investigation of genetically complex diseases.
同时考虑定性（患病状况）和定量（相关易感性指标）信息来定义双变量表型，可显著提高疾病风险估计的准确性和效率。本文提出了一种基于父母患病状况和子代定量性状计算子代发病风险的通用方法。我们还描述了多因素遗传的两种不同双变量模型，如计算机程序POINTER和YPOINT中所实现的，并明确了在估计计算发病风险所需的个体内及亲子相关性时它们的假设/约束条件。我们使用纽约高危项目中关于精神分裂症的心理测量家庭数据来估计这些相关性，并阐述我们的方法。我们的结果表明，即使一个性状仅与易感性中度相关，纳入定量性状信息也能解决一系列仅依靠父母患病状况无法得出的子代风险问题。双变量模型为在遗传复杂疾病研究中纳入易感性定量指标提供了一种有用的方法。
