It is of great interest to identify parent-of-origin effects (POEs) since POEs play an important role in many human heritable disorders and human early life growth and development. POE is sometimes referred to as imprinting effect in the literature. Compared with the standard logistic regression analyses, retrospective likelihood-based statistical methods are more powerful in identifying POEs when data are collected from related individuals retrospectively. However, none of existing retrospective-based methods can appropriately incorporate covariates that should be adjusted for if they are confounding factors. In this paper, a novel semiparametric statistical method, M-HAP, is developed to detect POEs by fully exploring available information from multilocus genotypes of case–control mother–child pairs and covariates. Some large sample properties are established for M-HAP. Finite sample properties of M-HAP are illustrated by extensive simulation studies and real data applications to the Jerusalem Perinatal Study and the Danish National Birth Cohort study, which confirm the desired superiority of M-HAP over some existing methods. M-HAP has been implemented in the updated R package CCMO.

中文翻译：

---

使用病例对照母子配对多位点基因型数据对起源效应的协变量调整推断

由于 POE 在许多人类遗传性疾病和人类早期生命成长和发育中发挥着重要作用，因此识别起源父母效应 (POE) 具有重要意义。POE 在文献中有时被称为印记效应。与标准逻辑回归分析相比，当从相关个人回顾性收集数据时，基于回顾性可能性的统计方法在识别 POE 方面更强大。然而，现有的基于回顾性的方法都不能适当地纳入协变量，如果它们是混杂因素，则应该对其进行调整。在本文中，开发了一种新的半参数统计方法 M-HAP，通过充分探索来自病例对照母子对和协变量的多位点基因型的可用信息来检测 POE。为 M-HAP 建立了一些大样本属性。M-HAP 的有限样本特性通过广泛的模拟研究和对耶路撒冷围产期研究和丹麦国家出生队列研究的真实数据应用进行了说明，这证实了 M-HAP 优于某些现有方法的预期优势。M-HAP 已在更新的 R 包 CCMO 中实施。