Abstract:
|
Gene set enrichment analysis (GSEA) has been applied to identify gene sets enriched by biologically "interesting" genes in many individual transcriptome studies. One limitation is that individual studies often have insufficient sample sizes to achieve reliable conclusions. With increasing data having been collected from various studies, several meta-analysis methods of gene set enrichment studies have been developed for analyzing gene-level expression. However, recent studies have indicated that analysis of isoform-level expression may lead to improved results over the analysis of gene-level expression. We propose meta-analysis methods to combine gene set information and isoform expression data from multiple RNA-seq studies, to enhance the efficiency of identifying enriched gene sets. Unlike existing meta-analysis methods that all use summary statistics, our approaches are model based, which integrate ideas from fixed-effect and random-effects methods. Besides, our methods can be applied to both discrete and continuous phenotypes including survival outcomes. Simulation and real data analysis will be conducted to compare the statistical power of our methods with existing methods.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.