Evaluation of procedures for amplification of small-size samples for hybridization on microarrays.

Various approaches have been developed for the preparation of samples for gene expression monitoring. For Affymetrix chips, a standard protocol is widely used; however, this is inefficient for small samples such as laser capture microdissections. Several amplification procedures for such samples already exist, and our goal was to test two of them: the first is based on random PCR amplification, and the second, linear amplification, involves performing the standard protocol twice. We analyzed a dilution of a commercially available mouse brain total RNA preparation and microdissections from mouse hippocampus and striatum. We evaluated the quality of microarray data by analyzing several chip parameters and performing multiple comparisons. At the biological level, brain microdissections prepared with either method gave similar expression results. At the technical level, analysis of the commercial sample showed that random PCR amplification is more reproducible, requires smaller RNA input, and generates cRNA of higher quality than linear amplification.