Dear All I am using the pamr.cv function from library pamr to cross-validate the nearest shrunken centroid classifier. Th following command is being used: NSC.results =pamr.cv(fit = NSC.trained, data = data) The output contains cols. containing the threshold values, no of genes and the total errors. The optimal threshold is chosen to be the threshold giving the minimum error rate. The command used for this is as follows: optthresh <- NSC.results$threshold[minindex] pamr.listgenes(NSC.trained, data, threshold=optthresh, genenames=TRUE) However, if the total number of genes at the opt. threshold is only one the following error mssg. is obtained: Error in dimnames(res) <- list(NULL, c("id", gnhdr, schdr)) : I have tried this with several data sets and the error persists. Hence, is this an error in the code or is PAM not supposed to work in case of a single gene surviving the optimal threshold. Thanks in advance! ALok -- Ph.D scholar Centre of Computational Biology and Bioinformatics School of Information Technology JNU New Delhi [[alternative HTML version deleted]]