Multiple Contrast Tests for Testing Dose-Response Relationship in a Pre-Clinical Microarray Experiment

• Activity Number: 472
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Setia Pramana*+ and Dan Lin and Ziv Shkedy
• Companies: I-BioStat, Hasselt University and I-BioStat, Hasselt University and I-BioStat, Hasselt University
• Keywords: Microarray ; Dose-response ; Multiple contrast tests ; Isotonic regression ; Gene expression ; Trend test

Abstract

Dose-response studies are the fundamental study in pharmaceutical industries to investigate the dose dependency of the response. In recent years dose-response studies have been integrated with microarray technology where several microarrays are available for a sequence of increasing dose levels and the response is gene expression. To identify genes with expression levels that change with studied dose levels, several tests for homogeneity of the means against order restricted alternatives in a microarray setting are available: the likelihood ratio test, Williams, Marcus, M and the modified M. We propose the implementation of Multiple Contrast Tests to these tests. The idea is that the order restricted alternative hypotheses can be decomposed in several elementary alternatives with particular patterns of equalities and inequalities. These contrasts are tested and the maximum contrast test is used. The inference is using its asymptotic distribution, permutations and the Significance Analysis of Microarrays. The power of these tests is compared using a simulation study to investigate the performance of these procedures. A dose-response study with 6 dose levels and 11,562 genes is used.