Search tips
Search criteria 


Logo of jbtJBT IndexAssociation Homepage
J Biomol Tech. 2010 September; 21(3 Suppl): S29.
PMCID: PMC2918114

Sample Labeling and Analysis Approaches for the Affymetrix Whole-Transcript Gene ST Array

M. Wang2 and P. White2
1Biomedical Genomics Core, The Research Institute at Nationwide Children's Hospital, Columbus, OH, United States;
2Department of Pediatrics, The Ohio State University, Columbus, OH, United States



The Affymetrix Gene 1.0 ST arrays provide an alternative platform for performing expression analysis than traditional 3' gene expression analysis arrays. These arrays offers several advantages, including updated content, lower cost, and probe sets designed to span the entire length of a given transcript. However, since its release in 2007, adoption of this platform has been limited and questions remain as to optimal approaches for sample labeling and subsequent data analysis. To address these issues we compared two labeling chemistries to generate the biotinylated sense DNA using the Microarray Quality Control Consortium (MAQC) total RNA titration samples: Stratagene Universal Reference RNA (A), Ambion Human Brain RNA (B) and the two titration samples (C and D, consisting of 3:1 and 1:3 ratios of A to B, respectively). Samples were labeled in triplicate, using either the Affymetrix GeneChip® WT Sense Target Labeling kit or the NuGEN WT-Ovation™ Pico RNA Amplification System, and hybridized to the Human Gene 1.0 ST Array. We next evaluated different data preprocessing approaches for background correction, normalization and probe set summarization, including the RMA, PLIER and dCHIP methods. Statistical analysis was performed using the LIMMA package in R, to generate lists of significantly differentially expressed genes (P < 0.001) and a fold change > 2. Dependent upon the preprocessing method used, the Affymetrix labeling chemistry consistently identified 30-50% more significantly differentially expressed genes than were detected with the NuGEN chemistry. Analysis of titration data showed that both chemistries had similar performance, and enabled us to determine optimal data preprocessing approaches. Our data demonstrate that even though the NuGEN chemistry provides a solution for sample labeling limited amounts of starting material, it results in significant data compression. However, higher quality data is obtained with the Affymetrix WT Sense Target Labeling kit if sufficient starting material (100 ng) is available.

Articles from Journal of Biomolecular Techniques : JBT are provided here courtesy of The Association of Biomolecular Resource Facilities