Post-Hybridization Quality Measures for Oligos in Genome-Wide Microarray Experiments

摘要

High-throughput microarray experiments produce vast amounts of data. Quality control methods for every step of such experiments are essential to ensure a high biological significance of the conclusions drawn from the data. This issue has been addressed for most steps of the typical microarray pipeline, but the quality of the oligonu-cleotide probes designed for microarrays has only been evaluated based on their a priori properties, such as sequence length or melting temperature predictions. We introduce new oligo quality measures that can be calculated using expression values collected in direct as well as indirect design experiments. Based on these measures, we propose combined oligo quality scores as a tool for assessing probe quality, optimizing array designs and data normalization strategies. We use simulated as well as biological data sets to evaluate these new quality scores. We show that the presented quality scores reliably identify high-quality probes. The set of best-quality probes converges with increasing number of arrays used for the calculation and the measures are robust with respect to the chosen normalization method.