Obtaining quality microarray data via image reconstruction

Abstract

This paper introduces a novel method for processing spotted microarray images, inspired from image reconstruction. Instead of the usual approach that focuses on the signal when removing the noise, the new method focuses on the noise itself, performing a type of interpolation. By recreating the image of the microarray slide, as it would have been with all the genes removed, the gene ratios can be calculated with more precision and less influence from outliers and other artefacts that would normally make the analysis of this data more difficult. The new technique is also beneficial, as it does not rely on the accurate fitting of a region to each gene, with its only requirement being an approximate coordinate. In experiments conducted the new method was tested against one of the mainstream methods of processing spotted microarray images. Our method is shown to produce much less variation in gene measurements. This evidence is supported by clustering results that show a marked improvement in accuracy. © Springer-Verlag Berlin Heidelberg 2003.