Adaptation of New Statistical Ideas for Medicine

新的医学统计理念的适应

• 批准号: 7211455
• 负责人: BRADLEY EFRON
• 金额: $ 21.65万
• 依托单位: STANFORD UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 1993
• 资助国家: 美国
• 起止时间: 1993-01-15 至 2010-01-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7211455
• 关键词: Data Analyses; Data Set; Development; Etiology; Functional Magnetic Resonance Imaging; Genes; Image; Medicine; Methodology; Methods; Proteins; Purpose; Spectrum Analysis; Time; Ursidae Family; Work; research study; size; theories

DESCRIPTION (provided by applicant): Recent advances in experimental methods, such as microarrays, fMRI imaging, and protein spectroscopy, have vastly increased the size and scope of biomedical data analyses. This has challenged traditional statistical theory, which was mainly developed to deal with small "one-at-a-time" problems. New statistical methodology is being invented, and the purpose here is to bring these developments to bear on massively parallel data sets such as those produced in microarray experiments. Two complementary aspects of statistical inference in large data sets will be attacked; identifying individually significant cases (e.g. genes), among hundreds or thousands of possibilities and identifying groups of cases that work together to influence disease etiology.

描述（由申请人提供）：实验方法的最新进展，如微阵列，功能磁共振成像和蛋白质光谱学，大大增加了生物医学数据分析的规模和范围。这对传统的统计理论提出了挑战，传统的统计理论主要是为了处理“一次一个”的小问题而开发的。新的统计学方法正在被发明，其目的是将这些发展应用于大规模并行数据集，例如微阵列实验中产生的数据集。大型数据集中统计推断的两个互补方面将受到攻击：在数百或数千种可能性中识别单个重要病例（例如基因），并识别共同影响疾病病因的病例组。