Assessing Statistical Significance in Microarray Experiments Using the Distance between Microarrays

• Activity Number: 415
• Type: Contributed
• Date/Time: Thursday, August 12, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Douglas Hayden*+ and David Schoenfeld
• Companies: Massachusetts General Hospital and Massachusetts General Hospital
• Keywords: microarray ; distance matrix

Abstract

In microarray cluster analyses it is often considered evidence of a group effect when gene chips from subjects in the same group cluster together. However, formal methods of inference are needed. We develop a method using the distance matrix to test equality of the gene expression pattern and variability between factor levels in an experiment. The within-level mean distance measures the within-level variability. The between-level mean distance measures the difference in the gene expression pattern. Permutation tests are used to assess the statistical significance of these measures. Analysis can also be based on the distance between linear contrasts of chips rather than individual chips. We apply the methodology to the comparison of two protocols for whole blood leukocyte RNA isolation. This work was supported by The Inflammation and Host Response to Injury Large Scale Collaborative Research Grant, NIGMS U54 GM-62119-02.