PURPOSE Recent progress in DNA microarray technologies allows researchers to perform genome-wide screening to detect differentially expressed genes. Under the paradigm of false discovery rate control, this paper presents sample size methods. METHODS The author considers the following two scenarios: 1) planning the sample size to keep the 'burden of testing' (defined as the expected number of genes that have to be tested before a true discovery can be made) below a certain level, and 2) given a fixed amount of budget, balancing the number of subjects to be recruited and the number of genes to be tested to maximize the total number of true discoveries. RESULTS The study calculates sample sizes to minimize the burden of testing or to maximize the number of discoveries. CONCLUSIONS The present approach to sample size calculation bears more direct relevance to gene-discovery studies.