decideTestsDGE

decideTestsDGE is a function in the limma package (Bioconductor) used to decide differential gene expression across multiple contrasts. It operates on results from a linear-model analysis of gene expression data, typically after fitting with lmFit and applying empirical Bayes moderation via eBayes. The function summarizes, for each gene, whether it shows a significant expression change in each contrast or comparison of interest.

The input for decideTestsDGE is usually an object that contains test statistics and p-values for multiple contrasts, such as an MArrayLM object produced by limma. The function applies user-specified significance criteria, commonly involving p-value or adjusted p-value thresholds, and returns a matrix with one row per gene and one column per contrast. The matrix entries are typically coded as -1, 0, or 1, indicating down-regulation, not significant, or up-regulation, respectively, for the corresponding gene and contrast. The exact handling of multiple testing and thresholding can be controlled via function arguments, allowing per-contrast or combined criteria.

decideTestsDGE is used to produce a concise, interpretable summary of differential expression results across several comparisons, facilitating downstream visualization and reporting. It complements the broader limma workflow by condensing complex, multi-contrast results into a compact decision table that can feed into heatmaps, Venn diagrams, or other summaries of differential expression patterns.

In short, decideTestsDGE helps researchers determine and organize differential gene expression decisions across multiple experimental contrasts within the limma framework.