The use of DNA microarrays has spanned numerous disciplines of life science research. Despite the volume of studies utilizing this technology, no consensus exists on basic issues such as the determination of significantly altered genes in a given experiment, often leading to either false-negative or false-positive data. In this report, we study the effect of dilution of biological alterations on the detection level of gene expression differences using cDNA microarrays. We propose that subtle alterations in transcript levels of genes below the 2-fold level should be considered when replicate hybridizations are performed, because these subtle gene expression changes may be due to a robust response in few cells. We measured the effect of dilution of gene expression and found that differences in gene expression between the two cell lines assayed (HaCaT and MCF-7) were detected even after a 20-fold dilution factor. These results better our understanding of biological alterations that comprise a relatively small percentage of an assayed organ and help in the interpretation of gene expression data.