III:Small: Overlapping Clustering Analysis of Biological Networks

该项目形成了一个跨学科的研究团队，整合了计算机科学家和生物医学科学家的专业知识，以解决蛋白质相互作用数据分析中的挑战性问题。具体来说，该项目开发了一种新的方法来检测重叠集群上出现的大量蛋白质-蛋白质相互作用的数据和validatesthe计算方法在酵母中。大量的蛋白质相互作用数据为我们提供了一个很好的机会，系统地分析一个大的生命系统的结构，也使我们能够了解基本原则，如必要性，遗传相互作用，功能，功能模块，蛋白质复合物，蛋白质相互作用网络中的功能模块的识别是非常有趣的，因为它们经常揭示蛋白质之间未知的功能联系，预测未知蛋白质的功能。蛋白质可以包括在一个或多个官能团中。因此，需要在蛋白质相互作用数据中识别重叠簇。本项目开发了一种独特的方法，将领域知识与蛋白质相互作用数据相结合，使数据更加可靠。它还开发了一种独特的方法来支持蛋白质相互作用数据的重叠模块分析，智能地将生物信息集成到模块分析过程中。该项目的另一个独特之处是计算方法与生物验证的紧密结合。通过将每个功能模块中的未知蛋白质与已知蛋白质相关联，我们可以表明这些蛋白质积极地为分配给模块的相应功能工作。该项目还可以在其他处理具有模块化网络属性的数据的领域中找到广泛的应用，例如web网络，社交网络和技术网络。http://www.cse.buffalo.edu/DBGROUP/PPI-networks/index.html

This project forms an interdisciplinary research team with integrated expertise of computer scientists and biomedical scientists to tacklethe challenging issues in analyzing protein interaction data. Specifically,the project develops a novel approach to detecting overlapping clusters onemerging large volume of protein-protein interaction data and validatesthe computational approaches in yeast. The vast amount of protein-proteininteraction data provides us with a good opportunity to systematicallyanalyze the structure of a large living system and also allows us tounderstand essential principles like essentiality, genetic interactions,functions, functional modules, protein complexes, and cellular pathways.The identification of functional modules in protein interaction networks isof great interest because they often reveal unknown functional ties betweenproteins and hence predict functions for unknown proteins. A protein maybe included in one or more functional groups. Therefore, overlapping clustersneed to be identified in protein interaction data. This project develops a unique method to integrate domain knowledge with the protein interactiondata so that the data will be more reliable. It also developsa unique method to support overlapping modularity analysis forprotein interaction data that intelligently integrates biologicalinformation into the modularity analysis process. Another uniqueaspect of this project is the tight integration of computational methods with biological verification. By associating unknown proteins with the known proteins within each functional module, we can suggest that those proteins positively work for the corresponding functions that are assigned to the modules. This project can also find broad applications in other areas which handle data with the modular network property, such as web network, social networks, and technological networks.For further information see the project web page:http://www.cse.buffalo.edu/DBGROUP/PPI-networks/index.html