Tvedebrink Torben, Morling Niels
Fredrik Bajers Vej 7G, DK-9220 Aalborg East, Denmark.
Frederik V's Vej 11, DK-2100 Copenhagen East, Denmark.
Sci Justice. 2015 Dec;55(6):408-14. doi: 10.1016/j.scijus.2015.07.001. Epub 2015 Aug 14.
The increase in the number of forensic genetic loci used for identification purposes results in infinitesimal random match probabilities. These probabilities are computed under assumptions made for rather simple population genetic models. Often, the forensic expert reports likelihood ratios, where the alternative hypothesis is assumed not to encompass close relatives. However, this approach implies that important factors present in real human populations are discarded. This approach may be very unfavourable to the defendant. In this paper, we discuss some important aspects concerning the closest familial relationship, i.e., identical (monozygotic) twins, when reporting the weight of evidence. This can be done even when the suspect has no knowledge of an identical twin or when official records hold no twin information about the suspect. The derived expressions are not original as several authors previously have published results accounting for close familial relationships. However, we revisit the discussion to increase the awareness among forensic genetic practitioners and include new information on medical and societal factors to assess the risk of not considering a monozygotic twin as the true perpetrator. If accounting for a monozygotic twin in the weight of evidence, it implies that the likelihood ratio is truncated at a maximal value depending on the prevalence of monozygotic twins and the societal efficiency of recognising a monozygotic twin. If a monozygotic twin is considered as an alternative proposition, then data relevant for the Danish society suggests that the threshold of likelihood ratios should approximately be between 150,000 and 2,000,000 in order to take the risk of an unrecognised identical, monozygotic twin into consideration. In other societies, the threshold of the likelihood ratio in crime cases may reach other, often lower, values depending on the recognition of monozygotic twins and the age of the suspect. In general, more strictly kept registries will imply larger thresholds on the likelihood ratio as the monozygotic twin explanation gets less probable.
用于身份识别目的的法医遗传位点数量增加，导致随机匹配概率极小。这些概率是在为相当简单的群体遗传模型所做的假设下计算出来的。通常，法医专家报告似然比，其中备择假设被假定不包括近亲。然而，这种方法意味着现实人类群体中存在的重要因素被摒弃了。这种方法可能对被告非常不利。在本文中，我们讨论了在报告证据权重时，与最亲近的家族关系，即同卵（单卵）双胞胎有关的一些重要方面。即使嫌疑人不知道自己有同卵双胞胎，或者官方记录中没有关于嫌疑人的双胞胎信息，也可以做到这一点。推导出来的表达式并非原创，因为之前已有几位作者发表了考虑近亲关系的结果。然而，我们重新进行讨论，以提高法医遗传学从业者的认识，并纳入有关医学和社会因素的新信息，以评估不将同卵双胞胎视为真正作案者的风险。如果在证据权重中考虑同卵双胞胎，这意味着似然比会根据同卵双胞胎的患病率和识别同卵双胞胎的社会效率被截断在一个最大值。如果将同卵双胞胎视为备择命题，那么与丹麦社会相关的数据表明，为了考虑未被识别的同卵双胞胎的风险，似然比的阈值应大致在150,000至2,000,000之间。在其他社会中，犯罪案件中似然比的阈值可能会因同卵双胞胎的识别情况和嫌疑人的年龄而达到其他值，通常会更低。一般来说，登记制度越严格，似然比的阈值就会越大，因为同卵双胞胎作案的解释可能性越小。
