An important problem in microarray experiments is the detection of genes that are differentially expressed in agiven mumber of classes. We consider a straightforward and easily implemented method for estimating the posterior probability that an individual gene is null. The problem can be...