Objectives: To estimate the system efficiency of uncle-nephew relationship identification by increasing STR markers and adding reference samples based on the test results of simulated data and real samples, so as to provide references for selecting the appropriate number of STRs and reference samples for uncle-nephew relationship identification.
Methods: Five common models of uncle-nephew relationship identification were constructed by adding different reference samples. In each model, the likelihood ratio (LR) for 10 000 pairs of uncle-nephew relationships and 10 000 pairs of unrelated individuals were simulated by detecting 19, 39 or 55 STRs, and the system efficiency at different thresholds was simulated. The samples of the Han population in Zhejiang were collected, and 55 autosomal STRs were obtained by using SiFaSTRTM 23plex kit, Goldeneye® DNA ID 22NC kit and AGCU 21+1 PCR amplification kit. When 19, 39 and 55 STRs were detected, the LR of each model and system efficiency under different thresholds were calculated and compared with the simulation results.
Results: Under the same detection system, the calculated results of simulated data and corresponding true samples were basically consistent. In the same model, there was a positive correlation between the system efficiency of uncle-nephew relationship identification and the number of STRs detected. Moreover, the system efficiency of introducing relatives was higher than identifying only two individuals. The order of preference for the introduction of relatives was the full sibling (or mother) of the uncle and the full sibling (or mother) of the nephew.
Conclusions: The system efficiency of uncle-nephew relationship identification could be improved by increasing the number of STRs and introducing known relatives, which would provide the basis for selecting the most appropriate detection system and reference individuals in actual cases.
目的: 基于模拟数据和真实样本的检测结果，探讨增加STR数量及引入不同参考样本对叔侄关系鉴定模型系统效能的影响，为叔侄关系鉴定选择适宜的STR数量和参考样本提供参考。方法: 针对引入不同的参考样本构建5种常见的叔侄关系鉴定模型，每种模型分别模拟10 000对叔侄关系与10 000对对应的无关个体在检测19、39、55个STR时的似然比及不同阈值下的系统效能。收集浙江汉族家系样本，使用SiFaSTRTM 23plex身份鉴定系统、Goldeneye® DNA身份鉴定系统22NC和AGCU 21+1 STR荧光检测试剂盒进行55个常染色体STR分型，计算每种模型在检测19、39、55个STR时的似然比及不同阈值下的系统效能，并与模拟结果进行比较。结果: 在同一检测系统下，模拟数据与对应的真实样本的计算结果基本一致。在同一模型中，叔侄关系鉴定的系统效能与检测STR数量呈正相关关系，且引入亲属时的系统效能均较仅对叔侄两个体进行鉴定时有不同程度的提高。引入亲属的优先顺序为叔的全同胞（或母亲）、侄的全同胞（或母亲）。结论: 增加检测STR数量和引入已知亲属均可提高叔侄关系鉴定的系统效能，为实际案件选择最适宜的检测系统及参考个体提供依据。.
Keywords: STR; autosome; forensic genetics; system efficiency; uncle-nephew relationship identification.