This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. Molecular prediction and personalization is the future direction of medicine. With the rapid development of microarray technology, it has become increasingly promising to identify novel biomarkers for the diagnosis and prognosis of human disease. However, current molecular classifiers for the prediction of clinical outcomes are not optimized. The long list of molecular markers can be reduced and the prediction accuracy can be further increased by using appropriate data mining algorithms. There exists an urgent need for a general feature selection scheme for identifying potential diagnostic and prognostic markers from high-throughput data. Furthermore, the development of a suitable methodology for elucidating the complex molecular interrelations in disease progression is critical. The extracted biomarker patterns can be used to predict clinical outcome for individual patients. The goal of this proposal is to test the hypothesis that a systems biology framework combining bioinformatic, genomic, proteomic, and clinical approaches and information enables the construction of clinically important molecular prediction models. Specifically, we will (1) develop a general feature selection scheme to identify novel biomarkers from microarray data;and (2) optimize a network model to construct molecular prediction models for individualized clinical decision-making. The long-term goals are twofold: 1) to develop a novel model framework for identifying important biomarkers that contain valuable information concerning the molecular mechanisms and therapeutic targets underlying disease, and 2) to make accurate predictions in individualized diagnosis, prognosis, and therapeutics. We anticipate that the proposed computational framework will fill gaps in current bioinformatics research.
Showing the most recent 10 out of 306 publications
