Project Abstract:

The availability of complete genomic sequences, combined with recent technological advances, has led to the development of high-throughput assays that probe cells at a genome-wide scale. Molecular networks and their components can be measured by such assays at multiple levels: gene expression measurements, protein-protein and protein-DNA interactions, chromatin structure, and protein quantities, localization, and modifications. These high-throughput genome-wide data offer us great promise to understanding diseases at the cellular level. This will lead to develop some new diagnosis and therapy methods, like gene diagnosis and therapy and so on. Clinical decision analysis is an effective tool for deriving optimal solutions in diagnostic, therapeutic, and prognostic management, relying on the underlying decision models, which capture the target clinical problems, objective evidences, encoding clinicians’ subjective judgments and patients’ preferences. The desired faithful model should represent the diseases at both clinical level and cellular level. However, constructing such decision model is a knowledge intensive task. It is arduous to manually process and integrate all the knowledge needed for model construction, especially in this post genomic era. Hence it is desirable to automatically extract knowledge required for decision model construction from multiple data sources: databases, knowledge bases, and even experimental data. We are going to design and develop a framework which can infer such hierarchical decision models from multiple sources.