Email updates

Keep up to date with the latest news and content from BMC Bioinformatics and BioMed Central.

Open Access Highly Accessed Software

Thawing Frozen Robust Multi-array Analysis (fRMA)

Matthew N McCall1* and Rafael A Irizarry2

Author Affiliations

1 Department of Biostatistics and Computational Biology, University of Rochester Medical Center, Rochester, NY, USA

2 Department of Biostatistics, Johns Hopkins University, 615 North Wolfe Street, Baltimore, MD, USA

For all author emails, please log on.

BMC Bioinformatics 2011, 12:369  doi:10.1186/1471-2105-12-369

Published: 16 September 2011

Abstract

Background

A novel method of microarray preprocessing - Frozen Robust Multi-array Analysis (fRMA) - has recently been developed. This algorithm allows the user to preprocess arrays individually while retaining the advantages of multi-array preprocessing methods. The frozen parameter estimates required by this algorithm are generated using a large database of publicly available arrays. Curation of such a database and creation of the frozen parameter estimates is time-consuming; therefore, fRMA has only been implemented on the most widely used Affymetrix platforms.

Results

We present an R package, frmaTools, that allows the user to quickly create his or her own frozen parameter vectors. We describe how this package fits into a preprocessing workflow and explore the size of the training dataset needed to generate reliable frozen parameter estimates. This is followed by a discussion of specific situations in which one might wish to create one's own fRMA implementation. For a few specific scenarios, we demonstrate that fRMA performs well even when a large database of arrays in unavailable.

Conclusions

By allowing the user to easily create his or her own fRMA implementation, the frmaTools package greatly increases the applicability of the fRMA algorithm. The frmaTools package is freely available as part of the Bioconductor project.