The “Window t test”: a simple and powerful approach to detect differentially expressed genes in microarray datasets

摘要: This work focuses on differential expression analysis of microarray datasets. One way to improve such statistical analyses is integrate biological information in the design these analyses. In this paper, we will use relationship between level gene and variability. Using information, propose from multiple genes get a better estimate individual variance, when small number replicates are available, increase power analysis. We describe strategy named “Window t test” that uses which share similar compute variance then incorporated classic test. The performances new method evaluated by comparison with widely-used methods for (the Student test, Regularized test (reg test), SAM, Limma, LPE Shrinkage t). each case tested, results obtained were at least equivalent best performing and, most cases, outperformed it. Moreover, Window relies very simple procedure requiring computing compared other designed