Kinship type determination for first-degree relationships using different patterns of allele sharing along the genome

There is no widely accepted and commonly used way to differentiate between different types of first-degree relationships, especially for low coverage, pseudohaploid data, which is common in ancient DNA (aDNA) studies. To address this question, we have developed a new, straightforward, rule-based method that makes use of the different patterns of allele sharing between sibling pairs and parent-offspring pairs along the genome. Our method is well-suited for most ancient DNA datasets, as it can be used on low coverage, pseudohaploid data, and is able to robustly determine the first-degree relationship type of two samples that share as low as 20,000 SNPs (TPR=94%, FPR=4%, for parent-offspring couples and TPR=96%, FPR=6% for siblings). We have tested our method using both simulated data for different numbers of overlapping SNPs between the simulated individuals of known relatedness, and published, real aDNA data. Using our new approach, we were able to determine the relationship type of three first-degree pairs from Aşıklı, which had not been settled before.目前尚无广泛认可且通用的方法可用于区分不同类型的一级亲属关系，针对古DNA（aDNA）研究中常见的低覆盖度伪单倍体数据而言，这一问题尤为突出。为解决这一问题，我们开发了一种全新且简便的基于规则的方法，该方法利用了全基因组范围内同胞对与亲子对之间差异化的等位基因共享模式。本方法适配绝大多数古DNA数据集，可用于低覆盖度伪单倍体数据，且能够在两个样本仅共享低至20000个单核苷酸多态性（Single Nucleotide Polymorphisms，SNPs）的情况下，稳健判定二者的一级亲属关系类型：亲子对的真阳性率（True Positive Rate, TPR）为94%、假阳性率（False Positive Rate, FPR）为4%，同胞对的真阳性率（TPR）为96%、假阳性率（FPR）为6%。我们分别通过两类数据对本方法进行了测试：一类是针对已知亲缘关系的模拟个体间不同重叠SNPs数量生成的模拟数据，另一类是已发表的真实古DNA数据。借助本新方法，我们成功明确了三对此前未被界定的来自阿什克利（Aşıklı）的一级亲属关系类型。