基于高密度单核苷酸多态性的共祖远亲缘关系预测算法准确性研究

Accuracy Research on The Distant Kinship Relationship Prediction by IBD Algorithm

目的研究构建基于共祖(identity-by-descent,IBD)片段算法预测远亲缘关系分析流程并评估预测准确性。方法采用高密度单核苷酸多态性(single nucleotide polymorphism,SNP)芯片对253份家系样本进行检测,研究基于IBD片段算法的分析流程进行两两个体间亲缘关系预测,评估预测准确性。随机减少SNP位点,评估位点数对算法预测准确性的影响。结果IBD片段算法预测1~7级亲缘关系平均置信区间准确率为94.72%,预测可信度为99.77%,6级及以上亲缘关系预测时出现假阴性。随着SNP数量减少,预测准确性会出现一定程度的下降。结论IBD片段算法可用于7级以内亲缘关系的预测,该算法在群体遗传学、法医遗传学等领域有重要应用价值。

Objective To construct an analysis process for identity-by-descent(IBD)algorithm to predict distant relatives and evaluate the prediction accuracy.Methods 253 family samples were detected by using high-density whole genome single nucleotide polymorphism(SNP)chip.IBD algorithm was used to predict the genetic relationship between pairs of individuals,and the prediction accuracy was evaluated.The number of SNPs was randomly reduced to evaluate the effect of SNP numbers on the accuracy of the algorithm prediction.Results Among 1-7th kinship degree,the average confidence interval accuracy of IBD algorithm was 94.72%,and the prediction credibility was 99.77%.The false negative of IBD algorithm was found when kinship degree is 6 or higher.When the number of SNPs decreases,the prediction accuracy will decline to a certain extent.Conclusion The IBD algorithm can accurately predict the genetic relationship within the seventh kinship degree,and it has important application value in population genetics,forensic genealogy inference and other fields.