ABSTRACT
This chapter discusses the design and analysis of relatively simple comparative experiments involving microarrays. Some of the discussion applies to all the most widely used kinds of microarrays, that is, radiolabeled cDNA arrays on nylon membranes, two-color, fluorescently labeled cDNA or long oligonucleotide arrays on glass slides, or single color, fluorescently labeled, high-density short oligonucleotide arrays on silicon chips. The main focus, however, is on two-color complementary deoxyribonucleic acid (cDNA) or long oligonucleotide arrays on glass slides because they present more challenging design and analysis problems than the other two kinds.
As subfields of statistics, the topics of design and analysis of microarray experiments are still in their infancy. Entirely satisfactory solutions to many simple problems still elude us, and the more complex problems will provide challenges to us for some time to come. Much of what we present in this chapter could be described as first pass attempts to deal with the deluge of data arriving at our doors. Questions come in a volume and at a pace that demands answers; we simply do not have the luxury of waiting until we have final solutions to problems before we get back to the biologists. A major aim of this chapter is to stimulate other statisticians to work with their local biologists on microarray experiments and to come up with better solutions to the common problems than the ones we present here.
