- 期刊
分析混淆基因型資料的半母數法
Semiparametric Methods for Analyzing Unphased Genotype Data
摘要
單套型(haplotype)是在單條染色體上某特定幾個單核苷酸多型性(single nucleotide polymorphism, SNP)基因所訂定的形式。遺傳關聯性研究時常以單套型作為基礎的遺傳因子，並探索單套型與數量表現型間之關係。然而，因為兩個同源單套型的組成無法區別，導致僅能蒐集到混淆基因型(unphased genotype)資料。本文以半母數線性迴歸模式(semiparametric linear regression model)描述單套型對數量表現型的影響，發展混淆基因型資料有關秩的推論。提供的重抽法可彈性地估計標準誤與建立迴歸參數的信賴區間。此外，利用Wald形式的統計量與計分(score)統計量來從事假設檢定，並由一些模擬研究來評估所建議方法的表現。
並列摘要
A haplotype is a specific pattern of particular single nucleotide polymorphism (SNP) alleles on a single chromosome. Genetic association studies often use haplotypes as basic genetic factors and explore the relation between haplotypes and quantitative phenotypes. However, only unphased genotype data can be collected because the combination of the two homologous haplotypes cannot be distinguished. This paper used the semiparametric linear regression model to study the effects of haplotypes on the quantitative phenotype and developed rank-based inferences for unphased genotype data. The resampling procedure offers the flexibility to estimate the standard error and confidence interval for regression parameters. Moreover, the Wald-type statistic and the score statistic were used to perform hypotheses test. Some simulation studies are presented to evaluate the performance of the proposed methods.