Comments on: Empirical Bayes for data integration

Abstract Rognon-Vael and Rossell (2025) introduce a method of incorporating summary information from previous studies to tackle the variable selection problem in high-dimensional linear regression. They focus their attention on the univariate case where the task is to identify the model that leads to improved inference and prediction. They fuse data from the study of interest with summary information from similar studies using empirical Bayes and discuss the consistency properties of their approach. They provide a computation scheme for their empirical Bayes approach and illustrate its application to gene expression data.