Alcoholism is the outcome of complex interactions between the environment and multiple gene loci, which may encode pre-existing susceptibility, or contribute to the neuroadaptations underlying the process of developing dependence. Because of this, the prospect of simultaneous, genome wide, high-throughput analysis of gene expression allowed by microarray technology has met with great expectations. The hope has been that new insights into pathogenesis of substance disorders will rapidly be gained, leading to identification of novel treatment targets. The usefulness of this approach as a discovery tool in addiction research will be critically reviewed here. In this article, we describe the evolution of our experimental approaches, from first generation Affymetrix expression arrays to present high-density arrays, and from the use of original Affymetrix software to more advanced analysis of the probe signal, and different statistical approaches to creating candidate gene lists. Further, we address some methodological issues critical to the study of brain samples by microarray technology. We also summarize findings from several expression profiling experiments involving different animal models of alcoholism. The accumulation of expression data from different animal models allows mining the database for patterns of overlap. Such second level analysis depends on the generation of uniform and reliable datasets.