Wang Jinliang
Institute of Zoology, Zoological Society of London, London, NW1 4RY, United Kingdom.
Heredity (Edinb). 2025 Apr;134(3-4):155-161. doi: 10.1038/s41437-024-00739-5. Epub 2025 Mar 27.
EMIBD9 is a computer programme implementing two likelihood methods for estimating the 9 condensed IBD coefficients, Δ = {Δ, Δ, …, Δ}, between a pair of individuals from their genotype data. Inbreeding coefficients of and relatedness (or kinship coefficient) between individuals are then calculated from the estimated Δ. One method is designed to apply to a small sample or a sample containing a high proportion of close relatives where allele frequencies and their powers or products are poorly estimated by assuming a large sample of non-inbred and unrelated individuals. It adopts an expectation maximisation (EM) algorithm to estimate both Δ and allele frequencies jointly and iteratively. The other method is designed to apply to a large sample of individuals containing few close relatives. It is fast because it, like all previous estimators, estimates Δ only and does not make iterative updates of allele frequencies by accounting for the inferred relatedness through the EM algorithm. EMIBD9 has both methods implemented for multiple computer platforms (Windows, Mac and Linux), and the Windows version has a GUI that facilitates data input and results visualisation. The GUI can also be used to simulate genotype data which are used to investigate factors affecting relatedness estimation accuracy, to optimise the experimental design of a relatedness study, and to compare the performance of different relatedness estimators.
EMIBD9是一个计算机程序,它实现了两种似然方法,用于根据一对个体的基因型数据估计9个压缩的同宗系数Δ = {Δ₁, Δ₂, …, Δ₉}。然后根据估计出的Δ计算个体之间的近交系数和亲缘关系(或亲属系数)。一种方法设计用于小样本或包含高比例近亲的样本,在假设大量非近亲且无亲缘关系个体的情况下,等位基因频率及其幂次或乘积估计不佳。它采用期望最大化(EM)算法来联合迭代估计Δ和等位基因频率。另一种方法设计用于包含少量近亲的大量个体样本。它速度快,因为与之前所有估计器一样,它仅估计Δ,且不通过EM算法考虑推断出的亲缘关系来对等位基因频率进行迭代更新。EMIBD9在多个计算机平台(Windows、Mac和Linux)上实现了这两种方法,并且Windows版本有一个图形用户界面(GUI),便于数据输入和结果可视化。该GUI还可用于模拟基因型数据,这些数据用于研究影响亲缘关系估计准确性的因素、优化亲缘关系研究的实验设计以及比较不同亲缘关系估计器的性能。