The hypothesis is that genes involved in the immature schwann cell and promyelinating state will be upregulated and genes that are involved in the myelnating state will be down regulated. Bioconductor limma package [1,2,3,4] was used to preprocess 6 samples from mouse tissue-cultured cells, which are hybridized to Mouse WG-6 version 2 Illumina Array. To remove batch effects, the complete dataset were normalized by RMA. Before doing statistical analysis, we filter the probe-sets by present/absent calls using the Wilcoxon signed rank-based algorithm. 18,420 genes out of 45,281 genes were kept and the remaining ones were removed as “Absent” to reduce false positive rate. We identified differentially expressed gene list by fitting linear models to the normalized expression values. The empirical Bayes shrinkage was applied to t-statistics using the limma package in Bioconductor. In general, selected genes have greater than 2.0 fold-change and less than 0.05 fdr adjusted p-value.
|Date made available||2010|