This project will build a comprehensive infrastructure for resolving cellular phenomena, from the molecular to the tissue level. The key underlying contribution of the project is the use of domains within biomolecules as primary units of functional description. This goal poses several intellectual challenges, which form the focus of the proposal: (i) building robust analytical, algorithmic, and software infrastructure for domain inference to complement existing (experimental) domain data; (ii) resolving existing molecular interaction data into domain interaction data; (iii) developing techniques for inferring modularity and functional coherence in domain interaction data; (iv) investigate implications of domain-based analysis on identification of disease markers, intervention, and experimentation.
The analysis, simulation and design of complex systems is often performed in terms of intuitively natural components. However analysis in such terms does not always provide the most powerful descriptor of complex systems, for purposes of prediction, design, or understanding. An example of this is the normal mode analysis of linear or linearizable complex systems. The normal mode is a mathematical abstraction, but is a powerful descriptor of real-life physical systems. Normal modes also define modules in the hierarchical and coarse-grained description of systems. Modularity is a hallmark of biology. Deciphering biological modules, studying how they combine in a quantitative manner to yield function, and describing their hierarchy in producing larger phenotypes, is likely to establish fundamental laws that govern living systems. Leveraging expertise in molecular and cellular biology and algorithm development, this collaboration of four institutions will focus on developing algorithms, tools, and methods for studying and designing biomolecular modules.
