复杂疾病的基因调控网络构建及调控机制研究

The rapid development of high-throughput technologies provides massive data sources for studying complex diseases,how to mining the complex regulatory relationship within disease cells from massive diseases data is currently a challenging bio-informatics problem. The research based on the complex disease gene expression data analysis,launching some relative studies in view of some scientific problems,such as the construction of gene regulation network of the complex diseases and regulation mechanism.Aimed at the problem of the missing gene expression data, giving a data reconstruction method based on sparse expression;In order to solve the problem of the quite lack data effectively, using multi-scale characteristic of wavelet analysis to design a cross-platform data integration method;According to the characteristics of the sparse network, comprehensively use the mutual information and energy function to build the large scale gene regulatory network.Combined with the regularization vector feature extraction method and the relationship between gene regulation networks to optimize the selected feature genes, finding the important pathogenic genes.Based on the pathogenic genes and its regulation informations, putting forward a method to recognize the pathogenic pathways based on the theory of the grey correlation analysis.Overall utilize of pathogenic genes and pathogenic pathways information,raising a pathogenic module identification model based on network structure conservative features,which provides scientific evidence from the molecular level to reveal the development mechanism of complex diseases and clinical diagnosis,treatment and drug design.

高通量技术的飞速发展为复杂疾病的研究提供了海量的数据来源，如何从海量疾病数据中挖掘出疾病细胞内复杂的调控关系是当前生物信息学所面临的挑战性问题。本项研究以复杂疾病的基因表达数据分析为基础，针对复杂疾病的基因调控网络构建及调控机制等科学问题开展相关研究：针对基因表达数据缺失的问题，给出一种基于稀疏表达的数据重建方法；为了有效解决数据相当缺乏的问题，利用小波分析的多尺度特性设计一种跨平台的数据整合方法；针对网络稀疏性的特点，综合利用互信息和能量函数构建大尺度基因调控网络；结合正则化向量特征提取方法和基因网络调控关系对所挑选的特征基因进行优化，发现重要的致病相关基因；基于致病基因及其调控信息，提出一种基于灰色关联分析理论的致病通路识别方法；综合利用致病基因和致病通路信息，提出基于网络结构保守特性的致病模块识别模型，为揭示复杂疾病的发生发展机制及临床诊断、治疗和药物设计提供了分子水平的科学依据。

高通量技术的飞速发展为复杂疾病的研究提供了海量的数据来源，如何从海量疾病数据中挖掘出疾病细胞内复杂的调控关系是当前生物信息学所面临的挑战性问题。本项研究以复杂疾病的基因表达数据分析为基础，针对复杂疾病的基因调控网络构建及调控机制等科学问题开展相关研究：针对基因表达数据缺失的问题，给出一种基于稀疏表达的数据重建方法；为了有效解决数据相当缺乏的问题，利用小波分析的多尺度特性设计一种跨平台的数据整合方法；针对网络稀疏性的特点，综合利用互信息和能量函数构建大尺度基因调控网络；结合正则化向量特征提取方法和基因网络调控关系对所挑选的特征基因进行优化，发现重要的致病相关基因；基于致病基因及其调控信息，提出一种基于灰色关联分析理论的致病通路识别方法；综合利用致病基因和致病通路信息，提出基于网络结构保守特性的致病模块识别模型，为揭示复杂疾病的发生发展机制及临床诊断、治疗和药物设计提供了分子水平的科学依据。