Adaptation of New Statistical Ideas for Medicine

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Biostatistical methodology is in a race to keep up with the massive databases being accumulated by new biomedical technologies. The long-term purpose of this grant has been to shorten the transfer time between modern statistical theory and its successful application to practice. The three new projects proposed here each use computer-intensive theoretical ideas to attack a difficult applications area: empirical Bayes methods for the analysis of microarray data; two way proportional hazards modeling for disentangling complicated survival situations that arise in cancer and HIV studies; and new permutation and bootstrap methods for assessing genotype-phenotype relationships in large association studies.