Currently the technologies used to image microarrays are detection systems for absorbance, fluorescence or luminescence quantification. Chargecoupled device- (CCD-) based non-confocal systems image fixed area sizes with a limited number of pixels over a scalable dwell time. Sensitivity increases with dwell time and number of photons per ^m2 detected. But the longer the dwell time the higher the dark current noise. Because of non-con-focality of CCD systems the intrascenic dynamic range is limited to functional 12- or 14-bit compared to 16-bit confocal laser scanning systems. The contrast from bright signal to background for the maximum resolution is 10-100 times better with microarray scanners compared to CCD imagers (Figure 16.1). The lower the number of fluorescence molecules the more sensitive the scanners are compared to CCD imaging systems for similar throughput. CCD chips acquire data on a limited area providing a snapshot depending on the size of the CCD chip. The data acquired by a CCD on a typical 3 x 1-inch microarray glass slide are stitched together. The background and signal correction is introduced by sophisticated software algorithms. This contributes to a lower image quality and reliability compared to images acquired by confocal scanners. The non-confocal detection of glass chips by a CCD camera results in a significant increase of background in the neighborhood of bright signals and the contribution of unspecific background from the backside of a contaminated glass chip. For transparent plastic chips and microplates the contribution of auto-fluorescence of the polymer is incredibly high for non-confocal imaging systems.

Scanners with an optimized confocality for planar glass surfaces currently provide the highest dynamic range and resolution in combination with sensitivity and speed:

(i) Laser scanning systems for gel detection (2D proteomics, 1D SAGE, etc.) have a lower optical resolution and are less confocal in order to optimize detection of larger features (mm range in all three dimensions) on larger areas (43 x 35-cm blots). Compared to microarray scanners, which are designed to scan 3 x 1-inch glass slides or even smaller chip areas to detect tens-of-thousands of 50- to 200-^m spots by 5- to 10-^m pixel digitization, typical gel scanners cannot be used for micro-array detection providing sufficient data quality.

(ii) Microarray laser scanners illuminate fluorescence markers of the sample pixel by pixel by moving the optics and/or the chip very fast (1030 Hz). The emitted fluorescence photons are gathered pixel by pixel in a photomultiplier tube (PMT) using mirrors, lenses, dicroics, filters, and pinholes. The PMT converts photons into electrons resulting in voltage-dependent analogue signal, which is translated by A/D converter into digital 16-bit raw image data. Depending on PMT gain

Figure 16.1.

Intrascenic Dynamic Range of images acquired by laser scanners is 10 times higher than for images acquired by CCD imagers.

(voltage-dependent) one photon generates a bunch of electrons translated into arbitrary counts of a pixel. Providing high signal-to-noise ratio on pixel, feature and image level as well as calibrating sample image data to a fluorescence standard is prerequisite to enable relative and quantitative comparable results.

0 0

Post a comment