PDF(Pubmed)
Abstract：
Human ear morphology prediction with SNP-based genotypes is growing in forensic DNA phenotyping and is scarcely explored in Pakistan as a part of EVCs (externally visible characteristics). The ear morphology prediction assays with 21 SNPs were assessed for their potential utility in forensic identification of population. The SNaPshot™ multiplex chemistries, capillary electrophoresis methods and GeneMapper™ software were used for obtaining genotypic data. A total of 33 ear phenotypes were categorized with digital photographs of 300 volunteers. SHEsis software was applied to make LD plot. Ordinal and multinomial logistic regression was implemented for association testing. Multinomial logistic regression was executed to construct the prediction model in 90% training and 10% testing subjects. Several influential SNPs for ear phenotypic variation were found in association testing. The model based on genetic markers predicted ear phenotypes with moderate to good predictive accuracies demonstrated with the area under curve (AUC), sensitivity and specificity of predicted phenotypes. As an additional EVC, the estimated ear phenotypic profiles have the possibility of determining the human ear morphology differences in unknown biological samples found in crimes that do not result in a criminal database hit. Furthermore, this can help in facial reconstruction and act as an investigational lead.
摘要：
基于SNP的基因型的人耳形态预测在法医DNA表型中正在增长，并且在巴基斯坦几乎没有探索作为EVC的一部分（外部可见特征）。评估了具有21个SNP的耳朵形态预测测定在人口法医鉴定中的潜在效用。SNaPshot™多重化学，使用毛细管电泳方法和GeneMapper™软件获得基因型数据。用300名志愿者的数码照片对总共33种耳朵表型进行了分类。应用SHEsis软件制作LD图。采用序数和多项逻辑回归进行关联检验。在90％的训练和10％的测试对象中执行多项逻辑回归以构建预测模型。在关联测试中发现了几种对耳朵表型变异有影响的SNP。基于遗传标记的模型预测了耳朵表型，具有中等到良好的预测准确性，用曲线下面积（AUC）证明，预测表型的敏感性和特异性。作为额外的EVC，估计的耳朵表型图谱有可能确定在犯罪中发现的未知生物样本中的人耳形态差异，而这些样本不会导致犯罪数据库命中。此外，这有助于面部重建,并作为研究线索.
