|
Overview
The role of imaging systems in microarray analysis is to gather
data about the level of hybridization of labelled probes to microarray
spots. There are two main types of imaging systems: charge-coupled
devices (CCD) and laser scanning systems. Laser scanning systems
will be considered in this summary.
Laser scanners have one or more (usually two) lasers, most often
of the HeNe variety (Helium/Neon), tuned to specific wavelengths
determined by the dyes that are to be detected. Cy3 and Cy5 are
the most common dyes used in probe labelling; the lasers for a Scanarray
3000 system (GSI Lumonics, Inc.) emit at 543nm (green) and 633nm
(red). These wavelengths are in the regions of highly efficient
excitation of their target dyes and negligibly efficient emission
of the dyes. The wavelengths with highest excitation frequencies
(556nm and 650nm for Cy3 and Cy5, respectively) are not used because
these are in a part of the spectrum where there is a high intensity
of emission from the dyes. Consequently, the ambient light from
the lasers would be interpreted by the signal processing component
of the scanner as a signal from the fluorophores.
The laser beam width defines the resolution of an image; most have
resolutions around 10µm. This small area is picked up by the
detector, a photomultiplier tube (PMT) and converted into an image
pixel. The pixel usually has a 16 bit bandwidth, giving a grayscale
of 0 to 65,535 as possible values. Most laser systems are confocal;
that is, they are set up to focus the brightest intensity of the
beam at the surface of the slide. The advantages of a confocal beam
are efficiency of fluorophore excitation and reduction of noise
from fluorescing contamination above and below the surface of the
slide. The disadvantages are a tendency to rapidly photobleach the
dyes, especially Cy5, and a sensitivity to small departures from
planarity in the slide surface.
Scanning of a slide is most often done for one dye at a time. Initially,
a low power, low resolution (50µm) scan is done to tune the
laser power and PMT collection efficiency. This is done to prevent
saturation of signal (i.e., too many high intensity pixels have
65,535 as a value) and to equalize the overall signal between the
two channels. After tuning these values, the high resolution scans
are done, and a pair of image files (tagged image format [TIFF]
or bitmap [BMP]) are stored for subsequent spot detection.
Qualitative considerations in fluorescence imaging are (1) pixel
size relative to spot size, (2) sensitivity of the system to a given
fluor concentration and detectivity of a given DNA concentration,
and (3) signal-to-background (i.e., everything that is not a fluorescing
spot).
Laser Scanning
of a Microarray
This protocol describes how to scan a hybridized slide using a
ScanArray 3000 scanner.
Starting the system. Shut down the scanning computer. Turn
on the scanner, then start the computer. Open the Scanarray software.
Two image windows will appear inside the Scanarray window, one for
each channel.
Insertion of slide for reading. Click on the Eject Slide
button. Insert the microarray slide array surface up into the slide
carriage that appears
Specification of array extent. Select Start from the
Acquire menu; the Acquire Image window will appear
(refer to Figure 1). Set the scan size by selecting the Custom
radio button. Enter 1.0mm as the X start position, 17.5mm as the
Y start position to scan. In the X box, enter 20mm (the width of
the array); in the Y box, enter 40mm (length of array).
Do the following for each channel. Complete both steps for channel
2 (the Cy5 channel), then repeat with channel 1 (Cy3). Always
start with channel 2 in scanning.
Initial scan. To assess if laser power and PMT settings
for the channel will yield overall intensities covering the range
of intensity values (0-65,535), perform a Quick Scan (50
micron resolution). Set laser power to 65% of maximum and PMT
voltage to 80% of maximum as an initial approximation. Check the
Quick Scan box, select the channel being scanned, and click
Acquire. Be sure that only a single channel box is checked;
otherwise both channels will be scanned. The image from the channel
will appear in one of the image windows as the scan is made.
Manual adjustment of scanner settings. Compare the overall
intensities of the spots, and depending on what you see, make
adjustments to the laser and PMT settings to compensate. For example,
if the overall intensity is moderately low, increase laser power
from 65% to 70% and PMT from 80% to 85%; if it is extremely low,
increase laser power to 75%-80% and PMT to 90%-95%. Adjustments
of this type will take some experience. Higher laser power will
tend to bleach the sample faster and higher PMT will tend to cause
background fluorescence to be picked up. Perform another Quick
Scan or two to obtain good intensity, using this method; try not
to do more than 2 or 3 overall for each channel. Cy5 bleaches
faster than Cy3, and it has a lot of variability in how many times
it can be scanned.
Adjustment of intensities between channels. Compare images
after the above adjustments are made to see if overall intensities
for the Cy3 and Cy5 images are about the same (do not look at individual
spots to make this judgement, unless they are carefully calibrated
control spots). If necessary, make an adjustment to the laser power
and/or PMT for one of the channels to equalize intensity, and do
a further Quick Scan.
High resolution scan. Select Start from the Acquire menu.
Check to see if the scan start position and scan widths are as specified
above. Set the laser power and PMT to the values found in the Manual
Adjustment step above, set resolution to 10 microns, check both
channel boxes, and uncheck the Quick Scan box. Select Acquire. Channel
2 will be scanned first, then channel 1. It takes about 10-15 minutes
for each scan to be performed. Images will appear in the two image
windows.
Finishing the scan. Save the two images as TIF or BMP files.
TIF files are in greyscale, BMP files have artificial colors assigned
according to a user defined palette. Quantitation of spots on the
images is done using the TIF files.
|
