基于小波包-FastICA在阿尔茨海默症中的应用及其生物学分析

摘要

The gene expression profiles of Alzheimer’s disease（AD） has the characteristics of high-dimensional,high noise,high redundancy,making the searching of specific genes a huge space and long time,and this reduce the excavation pretreatment of algorithm and its biological analysis.So it is necessary to denoising and dimensionality reduction for gene expressing profiles.This article firstly reduce the dimensionality of gene expression data using the method of wavelet packet transform-SAM,and the experiment results prove that the wavelet packet method could extract the useful information of gene expression profiles better.;and then do matrix factorization analysis to pretreatment data with the fast independent component analysis（FastICA） algorithm and then select specific genes based on independent component.The sample classification experiments show that the specific genes extracted by using FastICA have higher significance and improve the classification Accuracy for the samples.Simultaneously,this article presents how specific genes clustering and how they expressing in Alzheimer’s disease data set by gene enrichment analysis,and provide a favorable biology and medical pathology of AD.%阿尔茨海默症（Alzheimer’s disease,AD）基因表达谱数据具有高维性、高噪声、高冗余性等特点,使得AD特异性基因的搜索空间巨大,搜索算法时间长,降低了算法的挖掘性能及其生物学分析。因此对其基因表达谱数据进行去噪和降维预处理是十分必要的。文中首先利用小波包变换-SAM方法对数据进行降维去噪,实验结果证明了小波包方法能较好地提取基因表达谱有用信息;然后应用快速独立成分分析（FastICA）算法对预处理后的数据进行矩阵分解分析,并根据独立分量选取特异性基因。在此基础上的样本分类实验表明,FastICA提取的特异性基因具有较高的显著性,能够提高样本的分类结果。同时,通过所提取特异性基因的富集性分析,文中给出了这些基因在阿尔茨海默症数据集中聚类情况及其基因表达情况,为AD的生物学及医学病理分析提供有利的依据。