Commentson theAnalysisof Unbalanced Microarray Data

Motivation: Permutation testing is very popular for analyzing microarray data to identify differentially expressed genes; estimating false discovery rates is a very popular way to address the inherent multiple testing problem. However, combining these approachesmaybeproblematicwhen sample sizes are unequal. Results: With unbalanced data, permutation tests may not be suitable because they do not test the hypothesis of interest. In addition, permuta-tion tests can be biased. Using biased p-values to estimate the false discovery rate can produce unacceptable bias in those estimates. Results also show that the approach of pooling permutation null distributions across genes can pro-duce invalid p-values, since evennon-differentially-expressedgenescan have different permutation null distributions. We encourage researchers to use sta-tistics that have been shown to reliably discriminate differentially-expressed genes, but caution that associated p-values may be either invalid, or a less effective metric for discriminating differentially-expressed genes.