| Alzheimer’s disease(AD)is an irreversible neurodegenerative disease.Due to the increasing aging of China’s population,China has become the country with the largest number of Alzheimer’s patients in the world,which poses a great challenge to China’s current medical and health system.In recent years,due to the low success rate of new drug research and development for AD and the poor therapeutic effect of listed drugs,people began to seek new methods to study the pathogenesis of AD.From the perspective of systems biology,the occurrence and development of complex diseases usually involve the synergy between multiple pathogenic genes,which is often difficult to detect by traditional biological methods.However,with the rapid development of high-throughput technology,it has brought a large number of omics data containing rich biological information,which has brought a new dawn for the systematic and comprehensive study of the pathogenesis of AD.How to combine multi omics data and biomolecular network data for disease analysis and mining is one of the hotspots of current research.The work of this paper is mainly divided into the following three points:(1)This paper summarizes the mainstream methods of studying AD based on biomolecular networks at home and abroad,and introduces the concepts and characteristics of various omics data,biomolecular networks,biomarkers and their applications in AD.(2)Aiming at the problem that the molecular biomarkers identified by traditional methods are difficult to explain the pathogenesis behind AD,a network method based on graph Laplace regularization(DEGs＿GLReclipsoid FN)is proposed.This method introduces the information of protein-protein interaction network into gene expression profile data by graph Laplace regularization.Then,by analyzing the internal relationship between gene expression profile data and disease phenotype,we can identify a group of smallest gene subsets,which can distinguish disease phenotype to the greatest extent.The experimental results show that the biomarkers identified by this method are more interpretable and biological significance.(3)In order to identify the damaged functional subnets caused by some genetic variations,a multi-stage difference network analysis method(MSDNA)is proposed.Firstly,based on the protein-protein interaction network,the gene mutation data is transmitted through the network to solve the sparsity of the data,and then combined with the gene expression profile data to construct the gene regulation network in different disease stages of AD.Then,through the analysis and comparison with the gene regulation network of the control group,the differential regulation network of each stage is obtained.The purpose of this method is to find the damaged subnet with significant changes in the regulation mode between it and its neighbor genes due to a group of gene mutations according to the disease phenotype.The experimental results show that this method can improve the performance of disease diagnosis. |
