R package for a variational Bayes approach facilitating adaptive regularization of feature groups as used in 'Adaptive penalization in high-dimensional regression and classification with external covariates using variational Bayes'. The main function is graper, which fits the variational Bayes model given data X, a response y and a group annotation for each feature in annot.