Semiparametric Models for Large Scale-Biomedical Data

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): This.aim of this proposal is to develop novel statistical methodology to address issues in the analysis of large scale data from biomedical studies, especially the studies of tumors and virus diseases. The problems arising from the analysis of DNA micorarray, proteomic and longitudinal data will be carefully investigated. The proposal focuses on developing innovative semiparametric techniques for removing systematic biases in microarray experiments, selecting significantly expressed patterns of genes and proteins at different time points and under different experimental conditions, and efficiently assessing the covariate effects and predicting individual response trajectory for longitudinal studies. The strength and weakness of each proposed method will be critically scrutinized via theoretical investigations and simulation studies. Related software will be developed. Data sets from ongoing biological studies on cancer and virus diseases will be analyzed by using the newly developed statistical methods. This study allows biologists to more effectively remove the impact of experimental variations inherited in microarray experiments and permits biologists to reveal more meaningful scientific results with lower false discovery rates. It provides cutting-edge tools for biologists to understand biological processes, molecular functions and cellular activities. It introduces new tools for medical scientists to unveil how the risk factors affect individual disease over time. These will result in improved disease classification, diagnosis, prognosis, and drug design, among other pharmaceutical, therapeutic and medical goals.