Iterative gridding for automated microarray image analysis

High-throughput environments for DNA microarray analysis require automation in all aspects. While feature extraction from microarray images may utilize the knowledge of a pre-defined gene spot geometry of an array, automated grid generation presents a challenging task due to practical inaccuracies. Grid and meta grid parameters must be adjusted by correctly aligning superimposed grid templates to the corresponding gene spots. We introduce algorithmic methods to determine true gene spots, exclude artifacts and estimate optimal parameters for grid spacing and positioning. We propose iterative refinement of the results to allow for an accurate quantification in the subsequent analysis of DNA signals.