Krawczak M, Bockel B, Sandkuijl L, Thies U, Fenton I, Harper P S
Institute of Human Genetics, Göttingen, Germany.
Am J Hum Genet. 1991 Oct;49(4):735-45.
A combined logistic regression and life-table analysis is presented on age-at-onset data for Huntington disease. Covariates included in the analysis were sex of the at-risk individual, parental age at onset, and sex of transmitting parent. Parental age at onset and parental sex were found to be significant covariates for age at onset in the offspring, and the appropriate logistic regression functions are calculated by maximum likelihood methods. These regression functions permit a more precise evaluation of carrier risks and likelihoods than hitherto was possible by simple computational means. We further introduce a novel method to account for sibship correlations in the significance assessment, using log-likelihood differences between different models.
本文针对亨廷顿舞蹈病的发病年龄数据进行了逻辑回归和生命表分析相结合的研究。分析中纳入的协变量包括高危个体的性别、发病时父母的年龄以及传递父母的性别。研究发现,发病时父母的年龄和父母的性别是后代发病年龄的显著协变量,并通过最大似然法计算了合适的逻辑回归函数。与以往简单的计算方法相比,这些回归函数能更精确地评估携带者风险和可能性。我们还引入了一种新方法,利用不同模型之间的对数似然差异来考虑在显著性评估中的同胞相关性。