A complex system such as the cell can be modeled by a network, whose nodes represent components in the system (e.g., genes) and edges represent relationships between components (e.g., interactions). An interesting property that seems to be ubiquitous among all networks of real complex systems is community structure, which means a network can be decomposed into some relatively independent sub networks (communities). In biological networks, communities are closely related to functional modules and can be considered as the basic building blocks of the cell. These communities, together with the interactions among them, can provide significant insight into to the organizing principles and dynamic behavior of biological systems. Therefore, to automatically discover communities and to characterize their functions are essential steps for understanding complex biological systems. However, automatic community discovery and analysis is hindered by the large sizes of networks, the lack of well-characterized objective functions, and tremendous noise in biological network data. By combining recent theoretical and technical advances achieved in statistical physics and computer science, we plan to develop a computational toolkit for discovering and analyzing communities, and to apply these tools to systematically analyze a large number of biological networks.
Specific Aim 1 : Develop efficient, robust, and flexible algorithms for discovering community structures in biological networks.
Specific Aim 2 : Use the algorithms developed in Specific Aim 1 to systematically explore the organizing principles of protein-protein interaction networks.
Specific Aim 3 : Analyze the dynamic patterns of community structures and other topological properties of co-expression networks of human genes. The long term goal of the PI's research is to understand the relationship between the structure of biological networks and the dynamic behavior of complex biological systems, and to apply this knowledge in addressing a number of health-related applications such as disease classification and drug target selection. This research will develop the crucial theoretical infrastructure, necessary computational tools, and meaningful collaborations to achieve this goal. This grant will increase the PI's research productivity and peer-reviewed publications and will allow the PI to collect sufficient preliminary data to propose a future R01 application. This grant will also provide valuable research and educational opportunities for under-represented minority students. Many complex diseases, such as diabetes and cancer, involve multiple genes that interact with each other and form a network. To understand the complexity of such networks is the key for better prevention and more effective treatment of diseases. This research proposal aims at developing computer programs to analyze such networks more efficiently and effectively, which will ultimately lead to an improved understanding of complex diseases and more effective drugs.
Showing the most recent 10 out of 26 publications
