【摘要】 单核苷酸多态性(SNP)与复杂疾病之间的关联近来备受关注。SNP选择是在处理高维少样本的遗传数据时经常使用的方法,目的是从成千上万的SNP中选择出对遗传表型和疾病影响最大的SNP,从而完成数据降维。针对常用的信息SNP选择算法存在的未能充分考虑选择遗传数据内部的机理或者选择的SNP子集未能高度代表所有的SNP的信息的问题,提出改进蚁群算法(ACO)来选择SNP。基于临床数据的实验表明,ACO比现有的BPSO/STAMPA和BPSO/MLR方法具有更好的收敛性。实验分别在两个数据集上进行,论文提出的ACO方法的重构准确度提高了约10%～20%;实验充分说明论文方法在SNP的选择中具有较好的效果。更多还原

【Abstract】 The association between single nucleotide polymorphisms(SNPs) and complex diseases has recently received much attention. SNP selection is a method often used in the processing of high-dimensional and low-sample genetic data. The goal is to select the SNPs that have the greatest impact on genetic phenotype and disease from thousands of SNPs,thus completing data dimensionality reduction. Aiming at the problem that the commonly used information SNP selection algorithm fails to fully consider the mechanism of selecting genetic data or the selected SNP subset fails to highly represent the information of all SNPs,this paper proposes an improved ant colony algorithm(ACO)to select SNP,applying linkage disequilibrium to the ant colony algorithm. Experiments based on clinical data show that ACO has better convergence than existing BPSO/STAMPA and BPSO/MLR methods,and experiments are performed on two data sets respectively. The accuracy of ACO method reconstruction proposed in this paper is improved about 10% to 20%,experiments fully demonstrate that the method has a good effect in the selection of SNP.更多还原