Accurate and reliable image analysis is key to getting meaningful data from any microarray experiment. Typically, samples are labeled with red and green fluorescent dyes to be competitively hybridized to microarrays containing complementary probes. TIFF images of the microarray are then obtained for automated analysis.Unfortunately, current tools to quantify the relative amounts of biological activity have inherent limitations that are susceptible to erroneous results. Researchers at the University of Nebraska have devised an imaging technique that uses an imaging clustering algorithm to more accurately analyze microarray data. For each pixel in a gene spot image, the intensity values of each fluorescent labeled marker (ie Cy5 and Cy3) is determined as a vector, and the set of these values for each spot is analyzed as a set of two clusters. In short, the relative incidence of the binding substance is determined as a ratio of differences between corresponding indices of the representative pairs.