Differential expression was detected in an analogous manner as for translation efficiency, applying a moderated gene-by-gene t-test followed by p-value adjustment for multiple testing.
We corrected for multiple testing by applying a 0.05 false discovery rate (Benjamini and Hochberg 1995) to the P values associated with the moderated F statistics that resulted from our contrast matrix.
Finish by applying an overglaze.
We computed RMA expression data [ 18] from the raw data (CEL files) and subsequently applied a moderated t-statistic [ 20] to identify upregulated genes (data not shown).
Differentially expressed probes between Treg and Th were calculated by applying the moderated F-statistic [ 80], an empirical Bayes method, Cut-offs were set to pFDR < = 0.001 and fold change > = 2×.
The knockout targets of each TF (in Reimand et al.'s study) were those differentially expressed genes identified by applying the moderated eBayes t-test [ 12] to Hu et al.'s TFKMs [ 8].
Differentially expressed genes were identified by a moderated t-test [ 27] applying a critical FDR q-value of 0.005.
Further, we estimated the fold change and standard errors by fitting a linear model (using the lmFit function) for each gene given the groups of arrays and then compute a moderated t-statistics of differential expression by applying empirical Bayes at a level of 0.05.
This method takes both summarized point estimates and variation across the replicates into account to obtain a more robust ranking of biomolecules [for comparison, results on the mean-summarized replicates are generated additionally by applying the widely used empirical Bayes moderated t-statistic, here referred to as eBayes (Smyth, 2004)].
First, we calculated an un-moderated t-statistic for every probe set by applying Welch's t-test.
