基于支持向量机的单核苷酸多态性表型预测
首发时间:2009-04-27
摘要:大部分人类遗传疾病都和非同义性单核苷酸多态性引起的氨基酸突变相关。文章对使用支持向量机预测非同义性单核苷酸多态性表型的方法作了改进。结果表明,核函数的选择取决于输入向量的个数和序列特征值个数之间的关系,在输入向量个数大于序列的特征个数,但是相差不是很大的情况下,线性核函数在不失预测准确度的情况下能够较大提高运算速度。
For information in English, please click here
Predicting the phenotypic effects of non-synonymous single nucleotide polymorphisms by SVM
Abstract:Most of inherited human diseases are related to the mutation of amino acid which is caused by non-synonymous single nucleotide polymorphisms (nsSNPs).Here, we improve a method of predicting the phenotypic of nsSNPs by Support vector machine (SVM). It is proved that, the selection of kernel function depends on the relationship between the input vectors and the features of the sequences. The linear kernel function has a better performance in improving the calculation rate when the input vectors are more, but not too much than the features of protein sequences.
Keywords: SNPs SVM kernel function sequence features
基金:
论文图表:
引用
No.3173246064512407****
同行评议
共计0人参与
勘误表
基于支持向量机的单核苷酸多态性表型预测
评论
全部评论0/1000