New multilocus linkage disequilibrium measure for tag SNP selection
New multilocus linkage disequilibrium measure for tag SNP selection
复制标题
用于标签 SNP 选择的新多位点连锁不平衡测量
DOI：
10.1142/s0219720017500019
复制
发表时间：
2017-02-01
影响因子：
1
通讯作者：
Chen, Haowen
中科院分区：
文献类型：
--
作者：
Liao, Bo;Wang, Xiangjun;Chen, Haowen
目前已经提出了许多选择最佳标签单核苷酸多态性（SNP）集的方法。这些方法大多是基于连杆不平衡（LD）。经典的LD测量，如D'和r2，经常用于量化两个标记（成对）连锁不平衡之间的关系。尽管这些测量方法在许多应用中取得了成功，但它们不能用于测量多个标记之间的LD。这些LD测量需要从单倍型数据集中收集的等位基因频率信息。本文提出了一种基于信息论的基于多位点LD测度的snp聚类算法。然后，根据标签snp的数量、预测精度等进行优化，在每个聚类中选择标签snp。实验结果表明，该方法可以直接应用于HapMap项目收集的基因型数据集，从而节省了单倍分型的成本。更重要的是，该方法显著提高了标签SNP选择的效率和预测精度。
Numerous approaches have been proposed for selecting an optimal tag single-nucleotide polymorphism (SNP) set. Most of these approaches are based on linkage disequilibrium (LD). Classical LD measures, such as D' and r2, are frequently used to quantify the relationship between two marker (pairwise) linkage disequilibria. Despite of their successful use in many applications, these measures cannot be used to measure the LD between multiple- marker. These LD measures need information about the frequencies of alleles collected from haplotype dataset. In this study, a cluster algorithm is proposed to cluster SNPs according to multilocus LD measure which is based on information theory. After that, tag SNPs are selected in each cluster optimized by the number of tag SNPs, prediction accuracy and so on. The experimental results show that this new LD measure can be directly applied to genotype dataset collected from the HapMap project, so that it saves the cost of haplotyping. More importantly, the proposed method significantly improves the efficiency and prediction accuracy of tag SNP selection.