Improved detection of differentially expressed genes in microarray experiments through multiple scanning and image integration

Abstract

The variability of results in microarray technology is in part due to the fact that independent scans of a single hybridised microarray give spot images that are not quite the same. To solve this problem and turn it to our advantage, we introduced the approach of multiple scanning and of image integration of microarrays. To this end, we have developed specific software that creates a virtual image that statistically summarises a series of consecutive scans of a microarray. We provide evidence that the use of multiple imaging (i) enhances the detection of differentially expressed genes; (ii) increases the image homogeneity; and (iii) reveals false-positive results such as differentially expressed genes that are detected by a single scan but not confirmed by successive scanning replicates. The increase in the final number of differentially expressed genes detected in a microarray experiment with this approach is remarkable; 50% more for microarrays hybridised with targets labelled by reverse transcriptase, and 200% more for microarrays developed with the tyramide signal amplification (TSA) technique. The results have been confirmed by semi-quantitative RT-PCR tests.

Saturation and Quantization Reduction in Microarray Experiments using Two Scans at Different Sensitivities

“…This means that experiments that follow the standard methodology of Schena et al (1995) cannot take direct advantage of their work. Romualdi et al (2003) propose multi-scanning at image level seeking a higher contrast but not oriented to error reduction. A drawback of this method is the need of aligned images which actually are not easy to produce.…”

Microarray and EST database estimates of mRNA expression levels differ: The protein length versus expression curve for C. elegans

“…It is important to note in this context that whereas expression data measured with the microarray method arise from a single, large experiment, the ESTs used in the abundance within the EST database method arise from the entire database, which is constructed from many experiments done under different conditions and often examining different subsets of genes of interest [21]. For these reasons, it would appear that the possible biases arising from the microarray method are better controlled and mitigated [9] in current experiments than are the biases that arise from using the abundance within the EST database to estimate expression levels [21]. Both the microarray and the abundance within the EST database methods are, of course, intrinsically noisy.…”

High parallelism, portability, and broad accessibility

“…Fluorophores are generally used for this (labeled technique), nevertheless the disadvantages of this technology and new applications envisaged for genebased tests have boosted the research in the direction of finding alternative (non-labeled) techniques. The drawbacks of fluorescence-based measurements are due to the fact that errors are introduced both in the fluorescence signalreading process [Romualdi et al 2003] and the labeling of the sample [Churchill 1953]. Moreover, the current setup needed for scanning fluorescent arrays is not compatible with the development of portable systems.…”