Poster Sessions

BIOINFO-10

Yiwen He

Y. He, E. Asaki, P. Lemkin, J. Powell

Study on the RMA Value Changes and Its Impact on Analysis Results as the Pool of Microarrays Expands

For Affymetrix gene expression arrays, pre-processing methods such as Robust Multi-Chip Average (RMA) normalize data across the whole set of arrays. Because the expression values are dependent on the set of arrays included in the process, this creates a problem for a collection of data with an evolving set of microarrays. We used data from the Jak-Stat Prospector gene expression database, a collection of microarray experiments contributed by members of the Trans-NIH JAK-STAT Initiative. This on-going project is continually adding arrays, making it important to understand how the change in expression values affects earlier analysis results. For this study, we analyzed results at a number of snapshots as the pre-processing pool expanded from 56 to 152 microarrays.