EMT/BSSE: Discovery of Gene and Protein Expression Patterns and Networks

来自测序项目、微阵列研究、基因功能研究、蛋白质相互作用、比较基因组学、结构生物学等的大型数据库正在快速增长。这些数据具有多种形式，包括表达数据、网络/交互数据、本体、文本文档和原始图像/光谱数据。综合挖掘方法，结合这些庞大的公共资源库，是回答重要科学问题的关键组成部分。这项研究的重点是开发新的数据挖掘技术，对来自多个来源的复杂基因/蛋白质表达和相互作用数据集进行集成分析和挖掘。具体地说，本研究的目的是：1)通过布尔表达和关联聚类挖掘挖掘基因表达和调控的模式。2)挖掘新的蛋白质相互作用和网络模块/基序。3)在多个基因组范围的数据集上集成功能关系挖掘。鉴于这种复杂的、网络化的数据在社会网络、生物网络、语义网等不同领域中的激增，本研究中开发的方法和算法将广泛适用于其他重要领域。

Large-scale databases from sequencing projects, microarray studies, gene-function studies, protein-protein interactions, comparative genomics, structural biology, and so on are growing at rapid rates. These data come in diverse forms including expression data, network/interaction data, ontologies, text documents, and raw image/spectrum data. An integrated mining approach, combining these vast public repositories, is a crucial component in answering important scientific questions. This research is focusing on developing novel data mining techniques for integrated analysis and mining of complex gene/protein expression and interaction datasets taken from multiple sources. In particular, this research aims at: 1) Mining patterns of gene expression and regulation via Boolean expression and coherent cluster mining. 2) Mining novel protein interactions and network modules/motifs. 3) Integrated functional relationship mining over multiple genome-wide datasets. Given the proliferation of such complex, networked data in a variety of domains such as social networks, biological networks, semantic web, and so on, the methodology and algorithms being developed in this research will be widely applicable to other important areas.