Introduction

Microarray experiments, in which many thousands of signals are monitored concurrently, are becoming widely used for genome-scale measurements. For full integration of this technology into personalized medicine, users will need the ability to demonstrate the quality of their measurement results (1,2,3,4,5). Scanner performance, although not usually considered as a component of experimental variability or bias, has the potential to be a significant contributor to experimental results. Methods to ascertain scanner performance and qualify signal measurement in microarray experiments are under development at the National Institute of Standards and Technology (NIST). These methods are designed to track scanner performance over time and enable comparability of scanners. The method described in this paper uses a commercially available tool well characterized for use in a related technology, Raman spectroscopy, and available as a homogenous glass.

Development of a scanner qualification method requires a tool that is photostable and has adequate fluorescence signal intensity at wavelengths and instrument settings commonly used in microarray experiments. Previous studies have characterized a possible reference material composed of successive dilutions of the organic dyes cyanine 5 (Cy5) and cyanine 3 (Cy3) (6,7). As an alternative to these dyes that have been demonstrated to have limited stability, NIST Standard Reference Material (SRM) 2242, certified for y-axis Raman spectral intensity correction at 488 nm, 514 nm, and 532 nm excitation wavelengths, and known to be photostable at these wavelengths (8) (https://srmors.nist.gov/view_detail.cfm?srm=2242), was investigated. Although not certified for use in this capacity or at 635 nm, SRM 2242 exhibits the necessary photostability at the excitation wavelengths of 635 nm and 532 nm, allowing scanner signal and signal-to-noise ratio (S/N) monitoring. In the current study, the photostability of the SRM enabled tracking the instrument response day to day, confirming that changes observed in experimental arrays scanned were not due to changes in the scanner response.

While regulations by the United States Food and Drug Administration (U.S. FDA) for analytical methods are in place requiring evidence that a specific product will meet predefined specifications (www.fda.gov/CDER/GUIDANCE/pv.htm), validation for analytical instruments, referred to as instrument qualification, is increasingly acknowledged as a fundamental component of data quality. Instrument qualification underpins analytical method validation, system suitability tests, and quality control samples, according to a recent United States Pharmacopeia (USP) chapter, USP-NF <1058>, on analytical instrument qualification (9). Using a photostable material such as the fluorescent manganese glass of SRM 2242 enables scanner qualification when the glass is used to track signal stability over time, including time periods before and after experimental scans. Ideally, the microarray scanner should respond consistently from day to day, producing signal responses that match within a predefined uncertainty. SRM 2242, with its certification for use with laser powers greater than that typically used by microarray scanners, is photostable and not subject to the degradation observed with the organic dyes typically used with experimental microarrays. The stability of SRM 2242 facilitated assessment of the microarray scanner performance separately from the performance of the material, again something not possible with most organic dyes. With a history of signal measurements using a stable material, the user can estimate the uncertainty introduced by the scanning process and be assured that the instrument variability is minimal relative to the experimental variability. Additionally, the user can be assured that the scanner performance is “in control,” and that performance of the scanner today is similar, within limits, to previous uses.

Using SRM 2242, signal intensities and S/N were compared with those of Cy3 and Cy5, with scans taken on the same days as the SRM. The signal intensity and S/N of SRM 2242 were tracked over three different five-week periods on three different scanners from the same manufacturer, making scanner-to-scanner assessment possible. The stability of the signal intensity and S/N over the time periods studied is indicative of the utility of SRM 2242 for microarray scanner qualification.

Materials and Methods

A single piece of SRM 2242 borate matrix glass (NIST, Gaithersburg, MD, USA), which was manganesedoped (0.15 wt.% MnO₂), was used throughout the study. The SRM is 10.7 mm × 30.4 mm × 2.0 mm. A holder the size of a microscope slide, 25 mm × 75 mm × 1 mm, was made to hold the SRM in place in the scanner. The holder was constructed from a polymethylmethacrylate (PMMA) sheet with a rectangle the size of the SRM glass cut out of the middle, into which the SRM was press-fit in place. Scans of the SRM were made on the smooth side of the glass; the frosted side of the glass is certified for Raman spectral correction. The smooth side of the glass was chosen for scanner measurements due to the decreased noise relative to the frosted side.