In this paper, we develop a method to estimate the sample size required for a microarray study to obtain statistically meaningful results. This method is helpful in designing a study to explore the association of a clinical variable of interest with the expression of a large number of genes. This method was used to design the currently active infant brain tumor study with the objective of identifying genes that are associated with prognosis.
This article appears in Bioinformatics 2005. Other co-author is Cheng Cheng (St. Jude-Biostatistics).