Department of Biostatistics Seminar/Workshop Series
A parametric model to estimate the proportion from true null in microarray studies
Chang Yu, PhD
Assistant Professor, Department of Biostatistics, Vanderbilt University School of Medicine
Wednesday, October 14, 1:30-2:30pm, MRBIII Conference Room 1220
Intended Audience: Persons interested in applied statistics, statistical theory, epidemiology, health services research, clinical trials methodology, statistical computing, statistical graphics, R users or potential users
In microarray analysis, two-component mixture models are often used to estimate the proportion from true null. Among the methods summarized by Broberg (2005), only the Beta Uniform Mixture Model makes an explicit parametric assumption about the second component $h(p)$, the distribution of p-values under the alternative. This raises the question what parametric distributions should be used. The current work derives such distributions for a broad range of test statistics. We show a Beta distribution is derived only in the case that the distribution functions of the test statistics under the null and under the alternative differ by a Lehmann alternative through their survival functions, but not in general. We demonstrate our methods on real data sets and through simulations.