Department of Mathematical Sciences, Aalborg University, Skjernvej 4A, DK-9220 Aalborg East, Denmark.
Melbourne Integrative Genomics, School of BioSciences and School of Mathematics & Statistics, University of Melbourne, Building 184, Royal Parade, Parkville 3010, Victoria, Australia.
Forensic Sci Int Genet. 2019 Jan;38:48-53. doi: 10.1016/j.fsigen.2018.10.004. Epub 2018 Oct 10.
We recently introduced a new approach to the evaluation of weight of evidence (WoE) for Y-chromosome profiles. Rather than attempting to calculate match probabilities, which is particularly problematic for modern Y-profiles with high mutation rates, we proposed using simulation to describe the distribution of the number of males in the population with a matching Y-profile, both the unconditional distribution and conditional on a database frequency of the profile. Here we further validate the new approach by showing that our results are robust to assumptions about the allelic ladder and the founder haplotypes, and we extend the approach in two important directions. Firstly, forensic databases are not the only source of background data relevant to the evaluation of Y-profile evidence: in many cases the Y-profiles of one or more relatives of the accused are also available. To date it has been unclear how to use this additional information, but in our simulation-based approach its effect is readily incorporated. We describe this approach and illustrate how the WoE that a man was the source of an observed Y-profile changes when the Y-profiles of some of his male-line relatives are also available. Secondly, we extend our new approach to mixtures of Y-profiles from two or more males. Surprisingly, our simulation-based approach reveals that observing a 2-male mixture that includes an alleged contributor's profile is almost as strong evidence as observing a matching single-contributor evidence sample, and even 3-male and 4-male mixtures are only slightly weaker.
我们最近提出了一种新的方法来评估 Y 染色体谱的证据权重（WoE）。我们没有尝试计算匹配概率，因为现代 Y 谱具有高突变率，这特别成问题，而是提出使用模拟来描述具有匹配 Y 谱的人群中男性数量的分布，包括无条件分布和在数据库中该谱的频率条件下的分布。在这里，我们通过证明我们的结果对等位基因梯和创始人单倍型的假设具有稳健性，进一步验证了这种新方法，并将该方法扩展到两个重要方向。首先，法医数据库并不是评估 Y 谱证据相关背景数据的唯一来源：在许多情况下，被告的一个或多个亲属的 Y 谱也可用。迄今为止，尚不清楚如何使用此附加信息，但在我们基于模拟的方法中，很容易将其效果纳入其中。我们描述了这种方法，并说明了当一些男性亲属的 Y 谱也可用时，一个人是观察到的 Y 谱来源的 WoE 如何变化。其次，我们将我们的新方法扩展到来自两个或多个男性的 Y 谱混合物。令人惊讶的是，我们基于模拟的方法表明，观察到包含被指控贡献者的谱的 2 个男性混合物的证据几乎与观察到匹配的单个贡献者证据样本一样强，甚至 3 个男性和 4 个男性混合物也只是稍弱。
