The advent of large-scale gene expression technology in pharmaceutical drug discovery has precipitated an acute need for bioinformatics systems and methods for data analysis. These include methods for: data storage; integration of LIMs information; assessment of chip quality; normalization; metrics for the confidence of fold changes; integration with annotation information; and clustering approaches for the analysis of temporal and time series information. We present a suite of tools, called ‘gene expression computation and knowledge organization’ (GECKO), developed at the Hoechst-ARIAD Genomics Center, LLC, for the analysis of high-throughput (thousands of scans per year) gene expression data from both oligonucleotide (Affymetrix) and spotted cDNA (Molecular Dynamics) technology. As quality metrics are central to meaningful data interpretation, we also present a method, grounded in a Bayesian framework, for estimating the distribution of fold changes from inputs of experimental noise and show how this can be used to establish a ‘P-value’ to represent the statistical significance of fold change.