Microarray-Based RNA Profiling of Breast Cancer: Batch Effect Removal Improves Cross-Platform Consistency
Figure 4
Detection of differentially expressed genes between ER-positive and ER-negative breast cancer samples (a) and between luminal A and luminal B samples (b), respectively. Significance analysis of microarrays (SAM) algorithm was applied to the Agilent dataset and to the 29K dataset both with and without ComBat intraplatform batch-effect adjustment, respectively. The resulting gene lists were then ranked according to their statistics. The plot shows the proportion of shared top genes between Agilent versus 29K unadj. (blue), Agilent versus 29K intraplatform adj. (red), and Agilent versus 29K intraplatform adj. using only high-variance genes (green), respectively.
