Method-related vs. Biological Variability in cDNA Microarrays

Microarray studies have become a significant source of large biological datasets. Large changes in gene expression between two conditions can be identified, with high confidence, across many orders of magnitude of absolute expression. However, the identification of small changes of low-level expression with confidence is hampered by a variety of method-related factors and it is important to separate these different effects if we are to obtain biologically reasonable interpretations of the results. This paper presents a brief but detailed examination of some of these effects, in an attempt to separate method-related phenomena that are of no biological interest from biologically significant responses. These results are based on an ongoing study of ethanol exposure effects on rats where important biological effects are expected to be subtle. Issues considered here include dye-specific effects, method-related slide-to-slide variations, differences between duplicate spots on the same slide, and biological treatment effects.