It has already happened that laboratories have invested thousands of dollars and hours into one generation of microarray technology, and then have had to move on to the next generation of microarrays. After several experiments, it becomes apparent to them that results across the generations of microarrays are not identical. The question then arises of how to salvage the information from the first generation of microarrays. The concern is, of course, that they will have to treat the data sets of different microarray generations as distinct and incomparable and therefore they will lose "statistical power" in their classification or clustering studies. This concern is magnified when the earlier studies involved unique and irreplaceable tissue samples which may no longer be available to be studied with the newergeneration microarrays. We illustrate this challenge here with some of our own experience with the Affymetrix platform. We hasten to add that the challenges we have had in combining results across generations of this platform have their correlates on other platforms.
Was this article helpful?