Differential expression. AgiMicroRna wraps the linear model features implemented in the limma package. The target file is used to define the factors to be used in the linear model that we are going to fit to each gene using the processed data. To specify the linear model we need to define a design matrix (step 1). A contrast matrix (step 2) is also needed to obtain the estimates of the contrast of interest. Moderated statistics are then obtained and the results are stored in a specific R object (step 3). Finally, specific functions implemented in AgiMicroRna are used to extract the results into separate output files (step 4).