Kinship Index Variations among Populations and Thresholds for Familial Searching
Kinship Index Variations among Populations and Thresholds for Familial Searching
复制标题
DOI：
10.1371/journal.pone.0037474
复制
发表时间：
2012-05-16
期刊：
影响因子：
3.7
通讯作者：
Budowle, Bruce
中科院分区：
文献类型：
--
作者：
Ge, Jianye;Budowle, Bruce
目前的家族搜索策略主要是基于常染色体STR基因座，因为法医DNA数据库中的大多数罪犯档案不包含Y-STR或线粒体DNA数据。通常有两种家族检索方法，基于身份的方法或基于亲属关系指数的方法。基于KI的方法是分析上优越的上级方法，因为考虑了等位基因频率信息而不是单独的等位基因计数。然而，如果未知的法医特征可能归因于多个可能的相关人群，则应计算多个KI。一个重要的实际问题是选择KI阈值，以限制搜索中的候选人列表。一般有三种设定KI阈值的策略：（1）SWGDAM建议6;（2）最小KI >= KI阈值;（3）最大KI >= KI阈值。这些策略进行了评估和比较，使用模拟数据和经验数据。最低KI将倾向于更接近法医特征所属人群的KI。最小KI >= KI阈值的性能优于最大KI >= KI阈值。SWGDAM策略对于大型数据库的家族检索可能过于严格（例如，例如，在一个实施例中，100万或更多的配置文件），因为它的KI阈值取决于数据库的大小，大型数据库的KI阈值比小型数据库排除真正亲属的概率更高。最小KI >= KI阈值策略是一个更好的选择，因为它提供了根据预定数量的候选人或假阳性/阴性率调整KI阈值的灵活性。IBS和KI两者的联合使用不会显著降低在候选列表中包括真正亲属的机会，但确实提供了更高的家族搜索效率。
Current familial searching strategies are developed primarily based on autosomal STR loci, since most of the offender profiles in the forensic DNA databases do not contain Y-STR or mitochondrial DNA data. There are generally two familial searching methods, Identity-by-State (IBS) based methods or kinship index (KI) based methods. The KI based method is an analytically superior method because the allele frequency information is considered as opposed to solely allele counting. However, multiple KIs should be calculated if the unknown forensic profile may be attributed to multiple possible relevant populations. An important practical issue is the KI threshold to select for limiting the list of candidates from a search. There are generally three strategies of setting the KI threshold for familial searching: (1) SWGDAM recommendation 6; (2) minimum KI >= KI threshold; and (3) maximum KI >= KI threshold. These strategies were evaluated and compared by using both simulation data and empirical data. The minimum KI will tend to be closer to the KI appropriate for the population of which the forensic profile belongs. The minimum KI >= KI threshold performs better than the maximum KI >= KI threshold. The SWGDAM strategy may be too stringent for familial searching with large databases (e. g., 1 million or more profiles), because its KI thresholds depend on the database size and the KI thresholds of large databases have a higher probability to exclude true relatives than smaller databases. Minimum KI >= KI threshold strategy is a better option, as it provides the flexibility to adjust the KI threshold according to a pre-determined number of candidates or false positive/negative rates. Joint use of both IBS and KI does not significantly reduce the chance of including true relatives in a candidate list, but does provide a higher efficiency of familial searching.