Among the applications of DNA microarray expression data, identification of differentially expressed genes is widely used to search for genes involved in specific biological processes. Previous studies have shown that replication is necessary to reliably identify differentially expressed genes. In practice, however, further guidelines are required to provide references for experiment design, that is, to fulfill the needs for determining the replication number on a desired confidence level or the detection sensitivity of the microarrays. Here, we performed simulations to build a simple statistical table of general guidelines adapted specifically for identifying differentially expressed genes using replications of a microarray experiment, and we present an accompanying nonparametric method, replicated studentized-deviate detection (RSD). The analysis results of the TNF-α-stimulated cells showed that RSD is simple and robust, and that the resulting statistical table is plausible.
ASJC Scopus subject areas