The Impact of Photobleaching on Microarray Analysis

Abstract

DNA-Microarrays have become a potent technology for high-throughput analysis of genetic regulation. However, the wide dynamic range of signal intensities of fluorophore-based microarrays exceeds the dynamic range of a single array scan by far, thus limiting the key benefit of microarray technology: parallelization. The implementation of multi-scan techniques represents a promising approach to overcome these limitations. These techniques are, in turn, limited by the fluorophores' susceptibility to photobleaching when exposed to the scanner's laser light. In this paper the photobleaching characteristics of cyanine-3 and cyanine-5 as part of solid state DNA microarrays are studied. The effects of initial fluorophore intensity as well as laser scanner dependent variables such as the photomultiplier tube's voltage on bleaching and imaging are investigated. The resulting data is used to develop a model capable of simulating the expected degree of signal intensity reduction caused by photobleaching for each fluorophore individually, allowing for the removal of photobleaching-induced, systematic bias in multi-scan procedures. Single-scan applications also benefit as they rely on pre-scans to determine the optimal scanner settings. These findings constitute a step towards standardization of microarray experiments and analysis and may help to increase the lab-to-lab comparability of microarray experiment results.

Keywords: DNA; bioanalytics; bioinformatics; cyanine dye; fluorophore; microarray; photobleaching.

Figure 1

(A) Change in measured intensity of Cy5-labeled cDNA spots with increasing number of scans, depending on their initial intensity; and (B) change in measured intensity of Cy3-labeled cDNA spots with increasing number of scans, depending on their initial intensity.

Figure 1

(A) Change in measured intensity of Cy5-labeled cDNA spots with increasing number of scans, depending on their initial intensity; and (B) change in measured intensity of Cy3-labeled cDNA spots with increasing number of scans, depending on their initial intensity.

Figure 2

Change in measured intensity of Cy5-labeled cDNA spots of equal initial intensity with increasing number of scans, depending on the PMT voltage.

Figure 3

(A) Three-dimensional illustration of the final model of Cy5-photobleaching for V_PMT = 950 (B) Three-dimensional illustration of the final model of Cy3-photobleaching for V_PMT = 700.

Figure 4

(A) Cy5-data sets (y(I0)) with model-data (y_hat(I0)) at V_PMT = 750. R²: 0.994, 0.995, 0.998, 0.998, 0.999 (from lowest I0 to highest); and (B) Cy5-data sets (y(I0)) with fits (y_hat(I0)) at V_PMT = 500. R²: 0.994, 0.997, 0.991, 0.995, 0.984 (from lowest I0 to highest).