This article is part of the supplement: Selected papers from the Seventh Asia-Pacific Bioinformatics Conference (APBC 2009) .

Research

A statistical framework for integrating two microarray data sets in differential expression analysis

Yinglei Lai¹ , Sarah E Eckenrode² and Jin-Xiong She²

¹Department of Statistics and Biostatistics Center, The George Washington University, 2140 Pennsylvania Avenue, N.W., Washington, D.C. 20052, USA

²Center for Biotechnology and Genomic Medicine, Medical College of Georgia, 1120 15th street, CA4098, GA 30912, USA

author email corresponding author email

BMC Bioinformatics 2009, 10(Suppl 1):S23doi:10.1186/1471-2105-10-S1-S23

Published:	30 January 2009

Abstract

Background

Different microarray data sets can be collected for studying the same or similar diseases. We expect to achieve a more efficient analysis of differential expression if an efficient statistical method can be developed for integrating different microarray data sets. Although many statistical methods have been proposed for data integration, the genome-wide concordance of different data sets has not been well considered in the analysis.

Results

Before considering data integration, it is necessary to evaluate the genome-wide concordance so that misleading results can be avoided. Based on the test results, different subsequent actions are suggested. The evaluation of genome-wide concordance and the data integration can be achieved based on the normal distribution based mixture models.

Conclusion

The results from our simulation study suggest that misleading results can be generated if the genome-wide concordance issue is not appropriately considered. Our method provides a rigorous parametric solution. The results also show that our method is robust to certain model misspecification and is practically useful for the integrative analysis of differential expression.