An Orthogonal Transformed Aggregation Statistic to Test Gene Regulation with Prespecified Gene Sets or Pathways

• Activity Number: 238
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Shuyan (Sabrina) Wan*+ and Nanshi Sha and Peggy Wong
• Companies: Merck Research Laboratories and Columbia University and Merck Research Laboratories
• Keywords: Microarrays ; Translational Statistics ; Correlation ; Genomic biomarker ; Permutation

Abstract

With accumulated biological knowledge on gene pathways, researchers are interested in evaluating if a biological process with pre-specified gene sets is turned on or off with drug intervention. In such a clinical setting, two correlated hypotheses can be formed, namely "sensitivity" and "specificity" test to test if a biological process is regulated by itself or at a global level (Yu and Wan, 2009). Previously a principal component adjustment was proposed for the correlation between genes. Here we propose an orthogonal transformed aggregation statistic to better address the correlations. We show by simulation that the test is valid and powerful with appropriate individual-level statistics for various gene correlation structures and clinical designs (parallel or cross-over). Two real data applications are also presented. The results are very promising.