Zhang Yiming, Wu Yufeng
Department of Computer Science and Engineering, University of Connecticut, 371 Fairfield Way, Unit 4155, Storrs, CT 06269-4155, USA.
iScience. 2022 Jul 16;25(8):104768. doi: 10.1016/j.isci.2022.104768. eCollection 2022 Aug 19.
In this paper, we address a problem: can we perform ancestry inference for parents from one or more children's DNA samples? That is, suppose the parents' genomes consist of segments of different ancestry, and our goal is inferring parental ancestry and at the same time, calling parental genotypes from given children's genetic data. Such ancestry inference may provide insights into recent ancestors from children's genomes, and potentially has applications in understanding genetic traits. At present, there exists no method for this inference problem. We present parMix, a method based on hidden Markov model (HMM) that can jointly infer parental ancestry and call parental genotypes from data of a small number of children. Simulation results show that parMix performs well in practice. It can provide reasonably accurate parental inference given data from a small number (say three) of children. parMix becomes more accurate when data from more children are used.
在本文中，我们探讨一个问题：能否从一个或多个孩子的DNA样本中推断父母的血统？也就是说，假设父母的基因组由不同血统的片段组成，我们的目标是推断父母的血统，同时从给定孩子的遗传数据中确定父母的基因型。这种血统推断可以从孩子的基因组中洞察其近代祖先，并且在理解遗传特征方面可能具有应用价值。目前，针对这个推断问题还不存在任何方法。我们提出了parMix方法，这是一种基于隐马尔可夫模型（HMM）的方法，它可以从少数孩子的数据中联合推断父母的血统并确定父母的基因型。模拟结果表明，parMix在实际应用中表现良好。给定少数（比如三个）孩子的数据时，它能够提供合理准确的父母血统推断。使用更多孩子的数据时，parMix会变得更加准确。
Zotero 插件
