Quality assessment and principal component analysis when combining multiple microarray studies

Abstract

With the rapid accumulation of high-throughput genomic data, meta-analysis and information integration has become commonplace in biomedical research. To date, most inclusion/exclusion criteria for microarray meta-analysis are ad hoc and arbitrary. In this talk, an objective quality assessment tool will be presented to decide the inclusion of microarray studies for meta-analysis. We will also present two MetaPCA approaches that extends from PCA analysis in single study to a meta-analysis framework. The new MetaPCA method can be applied to improve visualization, clustering and classification.