Gene expression data: The technology and statistical analysis

The desire to view the simultaneous behavior of genes affected by a stimulus at the total genome level has brought the scientific world to a new place in history. It is now commonplace to have an experiment that investigates the expression of thousands of genes across treatments and time points. Biologists are quickly understanding that in order to make sense of these data and the variation that is inherent in the experimental process, statistical models need to be employed. This article presents important aspects of the two most common microarray technologies, the spotted array and the oligonucleotide array, for the purpose of identifying common and unique features of each technology and the data produced. Statistical models are suggested, and the statistical literature reviewed, in an attempt to bring some level of simplicity to the daunting task of analyzing these data. We include two examples, each based upon one of the different technologies, suggest a statistical model, and present the results of the analyses in hopes of providing both encouragement and guidance to readers wanting to become more involved in this exciting field known as genomics.