Resutik Peter, Aeschbacher Simon, Krützen Michael, Kratzer Adelgunde, Haas Cordula, Phillips Christopher, Arora Natasha
Zurich Institute of Forensic Medicine, University of Zurich, Switzerland.
Department of Evolutionary Biology and Environmental Studies, University of Zurich, Switzerland.
Forensic Sci Int Genet. 2023 May;64:102850. doi: 10.1016/j.fsigen.2023.102850. Epub 2023 Feb 25.
Biogeographical ancestry (BGA) inference from ancestry-informative markers (AIMs) has strong potential to support forensic investigations. Over the past two decades, several forensic panels composed of AIMs have been developed to predict ancestry at a continental scale. These panels typically comprise fewer than 200 AIMs and have been designed and tested with a limited set of populations. How well these panels recover patterns of genetic diversity relative to larger sets of markers, and how accurately they infer ancestry of individuals and populations not included in their design remains poorly understood. The lack of comparative studies addressing these aspects makes the selection of appropriate panels for forensic laboratories difficult. In this study, the model-based genetic clustering tool STRUCTURE was used to compare three popular forensic BGA panels: MAPlex, Precision ID Ancestry Panel (PIDAP), and VISAGE Basic Tool (VISAGE BT) relative to a genome-wide reference set of 10k SNPs. The genotypes for all these markers were obtained for a comprehensive set of 3957 individuals from 228 worldwide human populations. Our results indicate that at the broad continental scale (K=6) typically examined in forensic studies, all forensic panels produced similar genetic structure patterns compared to the reference set (G'≈90%) and had high classification performance across all regions (average AUC-PR > 97%). However, at K= 7 and K= 8, the forensic panels displayed some region-specific clustering deviations from the reference set, particularly in Europe and the region of East and South-East Asia, which may be attributed to differences in the design of the respective panels. Overall, the panel with the most consistent performance in all regions was VISAGE BT with an average weighted AUC̅ score of 96.26% across the three scales of geographical resolution investigated.
从祖先信息标记（AIM）推断生物地理祖先（BGA）在支持法医调查方面具有巨大潜力。在过去二十年中，已经开发了几个由AIM组成的法医检测板，用于在大陆范围内预测祖先。这些检测板通常包含少于200个AIM，并且是针对有限的一组人群进行设计和测试的。相对于更大的标记集，这些检测板恢复遗传多样性模式的能力如何，以及它们推断未包含在其设计中的个体和人群的祖先的准确性如何，仍然知之甚少。缺乏针对这些方面的比较研究使得为法医实验室选择合适的检测板变得困难。在本研究中，基于模型的遗传聚类工具STRUCTURE被用于比较三个流行的法医BGA检测板：MAPlex、Precision ID祖先检测板（PIDAP）和VISAGE基本工具（VISAGE BT），并与一个包含10k单核苷酸多态性（SNP）的全基因组参考集进行比较。从全球228个人类群体的3957名个体的综合样本中获得了所有这些标记的基因型。我们的结果表明，在法医研究中通常检测的广泛大陆尺度（K = 6）下，与参考集相比，所有法医检测板产生了相似的遗传结构模式（G'≈90%），并且在所有区域都具有较高的分类性能（平均AUC-PR > 97%）。然而，在K = 7和K = 8时，法医检测板显示出与参考集的一些区域特异性聚类偏差,特别是在欧洲以及东亚和东南亚地区，这可能归因于各个检测板设计的差异。总体而言，在所有区域中性能最一致的检测板是VISAGE BT，在研究的三个地理分辨率尺度上的平均加权AUC̅ 分数为96.26%。
