Table_2_Selection of Optimal Ancestry Informative Markers for Classification and Ancestry Proportion Estimation in Pigs.xlsx

Using small sets of ancestry informative markers (AIMs) constitutes a cost-effective method to accurately estimate the ancestry proportions of individuals. This study aimed to generate a small and effective number of AIMs from ∼60 K single nucleotide polymorphism (SNP) data of porcine and estimate three ancestry proportions [East China pig (ECHP), South China pig (SCHP), and European commercial pig (EUCP)] from Asian breeds and European domestic breeds. A total of 186 samples of 10 pure breeds were divided into three groups: ECHP, SCHP, and EUCP. Using these samples and a one-vs.-rest SVM classifier, we found that using only seven AIMs could completely separate the three groups. Subsequently, we utilized supervised ADMIXTURE to calculate ancestry proportions and found that the 129 AIMs performed well on ancestry estimates when pseudo admixed individuals were used. Furthermore, another 969 samples of 61 populations were applied to evaluate the performance of the 129 AIMs. We also observed that the 129 AIMs were highly correlated with estimates using ∼60 K SNP data for three ancestry components: ECHP (Pearson correlation coefficient (r) = 0.94), SCHP (r = 0.94), and EUCP (r = 0.99). Our results provided an example of using a small number of pig AIMs for classifications and estimating ancestry proportions with high accuracy and in a cost-effective manner.

运用小规模祖先信息标记（AIMs）构成了一种高效且经济的策略，用以精确估算个体的祖先比例。本研究旨在从约60,000个猪种单核苷酸多态性（SNP）数据中提炼出少量而有效的AIMs，并估算亚洲品种和欧洲家养品种中的三个祖先比例[华东猪（ECHP）、华南猪（SCHP）和欧洲商业猪（EUCP）]。总计186个来自10个纯种猪的样本被划分为三组：ECHP、SCHP和EUCP。通过利用这些样本和一对余分类支持向量机（SVM）分类器，我们发现仅使用七个AIMs即可完全区分这三组。随后，我们运用监督式ADMIXTURE算法计算祖先比例，并发现当使用假混合个体时，129个AIMs在祖先估计方面表现出色。此外，我们还将61个群体的另一969个样本应用于评估这129个AIMs的性能。我们还观察到，这129个AIMs与使用约60,000个SNP数据进行的三种祖先成分估计高度相关：ECHP（皮尔逊相关系数（r）= 0.94）、SCHP（r = 0.94）和EUCP（r = 0.99）。我们的研究结果提供了一个实例，展示了如何通过使用少量猪种AIMs以高精度且经济高效的方式进行分类和祖先比例的估算。