Rahim Fakher, Galehdari Hamid, Mohammadi-Asl Javad, Saki Najmaldin
Golestan Blv. Toxicology Research Center, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran.
Genet Res Int. 2013;2013:546909. doi: 10.1155/2013/546909. Epub 2013 Aug 13.
Aims. This review summarized all available evidence on the accuracy of SNP-based pathogenicity detection tools and introduced regression model based on functional scores, mutation score, and genomic variation degree. Materials and Methods. A comprehensive search was performed to find all mutations related to Crigler-Najjar syndrome. The pathogenicity prediction was done using SNP-based pathogenicity detection tools including SIFT, PHD-SNP, PolyPhen2, fathmm, Provean, and Mutpred. Overall, 59 different SNPs related to missense mutations in the UGT1A1 gene, were reviewed. Results. Comparing the diagnostic OR, our model showed high detection potential (diagnostic OR: 16.71, 95% CI: 3.38-82.69). The highest MCC and ACC belonged to our suggested model (46.8% and 73.3%), followed by SIFT (34.19% and 62.71%). The AUC analysis showed a significance overall performance of our suggested model compared to the selected SNP-based pathogenicity detection tool (P = 0.046). Conclusion. Our suggested model is comparable to the well-established SNP-based pathogenicity detection tools that can appropriately reflect the role of a disease-associated SNP in both local and global structures. Although the accuracy of our suggested model is not relatively high, the functional impact of the pathogenic mutations is highlighted at the protein level, which improves the understanding of the molecular basis of mutation pathogenesis.
目的。本综述总结了基于单核苷酸多态性（SNP）的致病性检测工具准确性的所有现有证据，并介绍了基于功能评分、突变评分和基因组变异程度的回归模型。材料与方法。进行全面检索以查找所有与克里格勒-纳贾尔综合征相关的突变。使用基于SNP的致病性检测工具进行致病性预测，包括SIFT、PHD-SNP、PolyPhen2、fathmm、Provean和Mutpred。总共对59个与UGT1A1基因错义突变相关的不同SNP进行了综述。结果。比较诊断比值比，我们的模型显示出较高的检测潜力（诊断比值比：16.71，95%置信区间：3.38 - 82.69）。最高的马修斯相关系数（MCC）和准确率（ACC）属于我们建议的模型（分别为46.8%和73.3%），其次是SIFT（分别为34.19%和62.71%）。曲线下面积（AUC）分析表明，与所选的基于SNP的致病性检测工具相比，我们建议的模型具有显著的总体性能（P = 0.046）。结论。我们建议的模型与成熟的基于SNP的致病性检测工具相当，能够在局部和全局结构中适当反映疾病相关SNP的作用。尽管我们建议的模型准确性不是特别高，但在蛋白质水平突出了致病突变的功能影响，这有助于增进对突变发病机制分子基础的理解。
