To create an electrophoregram, a solubilized sample is applied to a matrix, and an electrical potential is applied across the matrix. Because proteins or nucleic acids with different amino acid or nucleotide sequences each have a characteristic electrostatic charge and molecular size, components within the sample are separated by differences in the movement velocities with which they respond to the potential.

We visualize the separated components within the gel or blot using isotopic labels, stains (e.g. Coomassie Blue, silver, ethidium bromide), or various fluorescent, luminescent and other detection systems. In analyzing the gel/blot, position is measured to determine the approximate molecular mass of the molecule, and density and size of spots or bands are measured to determine the amount or relative purity of molecules.

Electrophoregrams can be evaluated by eye, using simple measurement tools. A basic strip or spot densitometer and a ruler can provide data that are adequate for some purposes. More detailed analyses require more sophisticated measurement tools. For example, we may need to know the molecular weights and proportional purities or abundances of various compounds distributed across a lane. In these sorts of quantitative applications, image analysis offers many advantages.

Point (e), above, requires some explanation. Electrophoretic technology is constantly evolving as new methods appear and as established methods are improved. Many technical advances result in specimens that have special illumination, sensing, or analysis requirements (see the Ultra-Low Light Imaging and Imaging Plate Readers booklets). A well designed image analyzer should allow us to analyze different types of gels/blots using a variety of scanners. It should not be necessary to move from one image analyzer to another as the type of analysis task changes.

For example, MCID™ could be used with laser densitometers or video cameras for autorads or stained gels. The same systems could be used with an imaging plate reader for improved analyses of isotopic specimens, or with scanning fluorescence imagers or low light cameras for direct imaging of fluorescence, chemiluminescence, or bioluminescence. Installation of an MCID system should allow the lab to avoid adding dedicated systems for each new technical advance or new application (Figure 1).

Figure 1: An illustration of the central role played by MCID in acquiring, analyzing, and documenting gel/blot specimens. Almost any specimen can be imaged, by selecting an appropriate input device. The images are quantified and archived on the MCID computer. Finally, MCID prints both numerical data and images on a selection of output devices.

Competent image analysis can yield reliable data from gels and blots (e.g. Duewer et al., 1995, Patton, 1995), but how are we to determine which of the many commercial analyzers is competent? The approach used with other complex instruments would be to examine the manufacturer's specifications (with some caution), and to locate the device in the published literature. Unfortunately, quantitative accuracy is not specified in an image analyzer's sales literature. Further, most gel/blot analyzers have yet to develop reputations based upon extensive publication records. Therefore, evaluation is not a simple task, and many gel analyzers are sold primarily because they appear easy to use. Ease of use is a major attribute, but so is the production of valid, replicable data.

In comparison with the image analyzers used in other life science applications, most gel/blot analyzers are very limited in the tasks they perform and in the devices they support. It is common to have different analysis software running on different devices. For example, one analysis package may have come with a scanning densitometer, and quite another with an imaging plate reader. Each of these limited analysis packages will take less time to learn than a more powerful and general-purpose system. On the other hand, laboratory staff need to learn multiple programs, and cross-validation of data obtained using different analysis programs is rarely performed.

There will always be a place for the simplest possible systems in qualitative applications, such as basic gel/blot viewing and image archiving. However, molecular biology laboratories will inevitably transit from a qualitative to a quantitative treatment of gel/blot data. As this transition happens, a precise and flexible image analysis system is usually required. Our image analyzers are designed for quantitative applications, and for flexibility.

For basic gel/blot visualization, select the simplest and lowest cost system that meets your needs. If quantitative analysis is important, particularly if the lab is using more than one detection technology (e.g. film, fluorescence, phosphor plate), then please include our products in your evaluation process. Our systems have a long heritage of measurement, and this emphasis upon quantification sets our instruments apart from most others.

Gel/Blot Analyzers as Analytical Instruments

The simplest task that an imaging system can perform is to display an image. The imaging system is used to archive the specimen in digital format, to control image acquisition devices (such as scanning densitometers) and printers, and to improve the visibility of features using simple imaging techniques (e.g. pseudocolor coding). This type of system is usually referred to as a gel documentation system.

Once we go beyond simply digitizing, displaying, and printing a specimen, two types of measurement are called for.

• Spatial measurement is used to determine the position and size (e.g. area) of a feature. The imaging system must be able to calibrate to one or more nonlinear distance scales, such as molecular weight.
• Densitometry is used to measure the relative amounts within specific bands or spots. Uncalibrated densitometry is rarely adequate, because values are affected by nonlinearities in system response. Calibration to optical density (OD) allows assumptions to be made (e.g. the Lambert/Beer law) regarding the relation between density or fluorescence intensity and concentration. Calibration in units of concentration, using external standards, provides the most accurate concentration measurements.

Capabilities for precise calibration are a major criterion in discriminating a gel/blot analyzer from a simple gel documentation system. Gel analyzers should calibrate, and should also allow a calibration to remain valid, even when specimen distortions or system nonlinearities are present. As an example, consider a Southern or Western blot. With undistorted specimens, spatial measurements are relatively straightforward. As long as an image analyzer can be calibrated to molecular weight, precise measurements are possible. However, since many specimens do exhibit distortions (such as smiling or warping), we need functions for the correction of spatial distortion. These types of functions are absent in a gel documentation system.

The issues involved in measuring densities are discussed in detail in the Densitometry and Cameras and Scanners chapters of our Fundamentals of Image Analysis booklet. Typical problems are the need to maintain good contrast transfer with narrow bands (Figure 2), the need for broad dynamic range, and the need for high sensitivity in fluorescence and luminescence applications. For valid measurement, it is critical that the camera or scanner is appropriate to the specimen characteristics. Allowing you to select from a broad variety of cameras and scanners is another way in which a high quality gel/blot analyzer differs from a gel documentation system (see next section).

Figure 2: Contrast transfer of a scanner or camera affects density values. Cooled CCD (high sensitivity, good rendition of contrast) and intensified CCD (medium sensitivity, poor rendition of contrast) cameras viewed through a Nikon Micro-Nikkor 55 mm lens equipped with an interference filter tuned for the emission spectrum of ethidium bromide. The ethidium-stained DNA gel was illuminated with a Spectroline (Spectronics Corp., Westbury, NY) TC356A illuminator (365 nm). Each camera was calibrated to a step wedge, extending beyond the range of intensities found in the specimen. The graph shows profiles of the same lane in the intensified CCD and cooled CCD images. The profiles are offset from each other, to allow better visualization of the data.

Note: the pair of peaks at 400 Kbase pairs. The cooled CCD shows these as being sharply separated. In contrast, the intensified CCD's poor rendition of contrast makes these peaks blend together, and also decreases the size of the sharp peak at 200 Kbase pairs. These data illustrate that component selection affects quantification. If we just needed to verify bands by their molecular weight, the ICCD might be more convenient to use than the cooled CCD. If densitometry is necessary, the cooled CCD would be a better alternative.

Almost any camera or scanner is adequate for imaging specimens that are well illuminated, and that do not exceed about 1 optical density units (D) in range. However, not all specimens have these ideal characteristics. Fluorescent and luminescent gels/blots tend to have both very dim and very bright regions, that can span more than four orders of magnitude. Similarly, many opaque or stained gels/blots, and autorads, contain very dark regions that go well beyond 1D. The image acquisition device must be selected with these specimen characteristics in mind.

A gel/blot analyzer should allow the use of video cameras for routine work, and scanning densitometers, imaging plate readers, cooled cameras, and other devices when required. High precision (e.g. 12-bits, 16-bits) is necessary if small density differences are to be discriminated across a wide dynamic range. This is often the case with dense autorads, or with luminescence and fluorescence.

Strip Densitometers

One-dimensional gels have traditionally been analyzed with a simple strip densitometer. This instrument proceeds down a lane, passing density and position values to a plotter or a computer. The result is a graph of density variations across the lane. The better strip densitometers are sensitive (can detect small density differences), and have a wide dynamic range (can scan dense specimens). However, they are slow, read only one lane at a time, and generate a graph, not an image of the gel. As the graph is not superimposed over the actual specimen, it can be difficult to correlate the graphical display with the specimen.

Scanning Densitometers

Scanning densitometers move a collimated light beam or laser beam over the specimen. A detector measures the amount of light transmitted or reflected, and passes this value to a computer which assembles the discrete points into a complete image.

Scanning densitometers have a broad dynamic range and exhibit excellent densitometric precision (because they are internally calibrated, and measure both incident and transmitted or reflected illumination). Therefore, they are useful for opaque gels (such as membrane blots) containing very dense regions. However, they are slow. A scan of a standard small gel might easily take 10 minutes.

A somewhat less precise instrument is the linear scanning densitometer. This moves a linear illuminator and a linear array of detectors over the specimen. Linear scanning densitometers are faster than point-by-point by point scanners, and the familiar desktop scanners are very cheap. However, linear scanners have a narrower dynamic range than instruments which only illuminate and read from one part of the image at at time.

Scanning densitomters produce TIFF files, or files in proprietary formats (Molecular Dynamics, BioRad instruments). MCID/AIS accept these file formats, and allow you to analyze data from any number of different devices on a single image analysis system.

Standard Video Cameras

Standard video cameras are familiar to anyone who has set up a low-cost gel/blot imaging system. These cameras allow you to move the specimen about, adjust illumination, optical magnification, and so forth, all while viewing a digital image. The ability to digitize images so rapidly that they seem like broadcast TV is sometimes referred to as "real time imaging". It is a major convenience.

The major limitation of video cameras (and most other cameras for that matter) is a narrow range of response. For example, your eye would see a difference between a band at 2D (1% transmission) and another at 3D (0.1% transmission). This level of difference (less than 1% of light absorbed) is often meaningful in real specimens. Smaller differences would be seen accurately by a laser densitometer. In contrast, a camera would probably fail to detect any difference between 2D and 3D. Cameras are poor at discriminating small differences in density above about 1D.

We have settled on three standard video cameras for routine work. The Sony XC-77 is very low in cost, and offers excellent performance for autorads. The Dage 72 uses the same camera head as the Sony, but provides very flexible control of camera response from an external control unit, and has the ability to integrate for use with moderately bright fluorescence. The Dage 72S lacks the external control unit, but includes the integration capability.

Integrating Cameras

An integrating camera accumulates signal onto the CCD for a period of time. Integrating video cameras (like the Dage 72 mentioned above) accumulate signal over a number of frames. You might integrate two or more video frames (each being 1/30 sec), before an image is output from the camera. Asynchronous integrating cameras do not have fixed frame rates, and accumulate signal over any time period. We might integrate for 100 msec, or 1000 msec, or any other time interval. Many integrating cameras are chilled or cooled (chilled is above 0 degrees C, cooled is below) to improve their sensitivity.

Most gel documentation systems support integrating video cameras. A minority of systems also support high-quality digital integrating cameras. These high quality cameras are much more sensitive, and allow use of the system with much dimmer specimens. For a discussion of imaging at low light levels, see the Ultra-Low Light Imaging booklet.

Imaging Plate Readers

Film is an inefficient recorder of particle emission. For example, less than 5% of the beta particles emitted by a 32P-labeled sample actually interact with the emulsion of a standard X-ray film to form an image. Therefore, long exposures may be required. Film will also fail to record any differences in labeling larger than about 300:1. The actual usable range is much less. The linear response range of film (the useful part) occupies only a portion (about 100:1) of its full exposure range.

Photostimulable storage phosphor plates are widely used as an alternative to film for recording and quantifying autoradiographic images. The advantages of storage phosphors are a linear dynamic range covering five orders of magnitude, and a much faster exposure than film. For a detailed discussion of this technology, see the Imaging Plate Readers booklet.

The software supplied with imaging plate readers is usually primitive. AIS and MCID provide an excellent alternative to the standard software, and act as a host/controller and analysis station for the Fuji BAS family of phosphor plate imagers. MCID also accepts images made on the MD, BioRad, and Packard imaging plate readers, using the native image formats of these devices. That is, the images are imported into MCID at full precision and with all internal calibrations intact.

Fluorescence Imagers, Multipurpose Imagers, Etc.

As alternatives to isotopic labeling, a growing variety of fluorescence and luminescence-based scanners is becoming available. Given the new scanner types and the rapid expansion of luminescence and fluorescence techniques, it is difficult to know exactly what types of files a laboratory may have to analyze in the future. Therefore, MCID and AIS offer two ways to analyze data from almost any imaging device.

Import the file in TIFF format. Most image formation devices can create files in TIFF format, though some lose precision during the conversion process (e.g. a 16-bit image is converted to 8-bit TIFF).

Use our image template editor to create an import filter for the device's native file format. In this case, the full precision of the image file is retained.

Over the years, we have scanned specimens with almost every type of detector. This experience leads us to suggest some general guidelines.

Brightfield Illumination

Typical images include film autoradiographs, stained protein gels, and film images of luminescence. For all of these purposes, cameras are convenient, fast, and are useful for specimens which contain diffuse densities of up to about 1.2-1.4.

Unfortunately, cameras have limitations in scanning dense specimens. Very dense peaks or spots will be read with greater amounts of error, a particular problem for membrane assays. Beyond about 1.4D, cameras are not suitable for image acquisition. Laser and other scanning densitometers are slow and tedious to use, but are accurate across a wide dynamic range (>3 OD units, Figure 3).

Figure 3: Coomassie-stained gel scanned on a Molecular Dynamics Personal Densitometer, and analyzed at 12 bit precision (4096 gray levels) using an MCID M2. Above we see the wide range of densities present in this specimen (up to about 2.3 D). Below, we zoom in on the range below 0.2 D, to show information in the lightest portions of the gel. Because the laser densitometer has wide dynamic range and 12 bit precision, both the dense and very light data can be quantified. We could accurately read the primary peak, and each of the smaller peaks. A camera scan of the same specimen underestimated the very dense primary peak.

Standard document scanners (e.g. the Hewlett Packard ScanJet) are adequate for qualitative analyses, and for reading positional information. They can read gels/blots over a broader dynamic range than cameras (Figure 4), but have limited precision within the more dense parts of the specimen. We have not made a careful evaluation of this, but better results are available from some high-precision scanners that have been developed specifically for scanning medical radiographs and biological specimens.

Figure 4: Calibration of a Hewlett Packard ScanJet 3c (with transparency adapter) to a Kodak diffuse density step wedge. The scanner was operated in linear mode, with no brightness or contrast adjustment. Each point on the step wedge is shown as a dot, and a regression line is plotted. The scanner is fairly linear across a range of about 2 OD units. However, its 8 bit (256 gray level) density scale is inadequate for accurate densitometry across this broad dynamic range.

Document scanners can also be used to image anatomical specimens, such as autorads of radiolabeled tissue sections. However, the images are generally inferior to those made with good quality cameras (Figure 5), and cameras are more convenient to use.

Figure 5: Comparison of rat brain autoradiographs scanned with a video camera, and a Hewlett Packard ScanJet 3C document scanner (scanning at 1200 dpi) with transparency adapter. You should be able to see, even in this printed document, that the video image is sharper and has a more realistic rendering of gray shades.

Video Camera

1200 dpi scanner

Fluorescence

Non-cooled Cameras

Low-cost, non-cooled integrating cameras can acquire moderately bright fluorescent specimens (such as EtBr gels), but they do exhibit a limited dynamic range. If the camera is adjusted to image the majority of bands in an ethidium-stained gel, the brightest bands can exhibit severe flare, while fainter bands may not be visible. This may or may not be a problem for your application. If it is a problem, a poor solution is to use the camera in a nonlinear mode (compressing response to bright peaks). Compression can help in preventing flare, but it does not allow accurate densitometry.

Some high-quality non-cooled integrating cameras (e.g. MegaPlus, Xillix) are more sensitive than standard integrating video cameras. They achieve relatively high precision (typically 10 bits) within a dynamic range that is wide enough to be useful for most moderate to bright fluorescent materials (Figure 6).

Figure 6: Ethidium bromide-stained DNA blot illuminated at 365 nm, and acquired at 10 bit (1024 gray levels) precision using a Xillix MicroImager camera. This graph (from an MCID M2) zooms in on a region containing small DNA fragments, and displays faint bands (below 75 gray levels). In other parts of the blot, labeling was much brighter with fluorescence values up to about 1000 gray levels. Note the peak at about 0.6 kbase pairs. This very faint peak is only about 10 gray levels above baseline. It can be quantified easily, even though it occupies only about 1% of the full range of fluorescent intensities in the gel.

Chilled and Cooled Cameras

Chilled CCD cameras are those which cool the CCD chip to temperatures above 0 degrees C. Typical cameras of this type include the Photometrics Sensys, and the Hamamatsu C5985. Cooled cameras are those which cool the CCD element to below zero.

Chilled cameras generally offer 8-10 bit precision, and are very flexible. They are suitable for EtBr gels, for example, and can also perform quantitative autoradiography, fluorescence microscopy, and other tasks.

Cooled cameras have higher sensitivity than chilled cameras. To take advantage of the high sensitivity, they tend to offer high precision digitizers (e.g. 12-16 bits). A cooled, high-precision camera is excellent, but probably unnecessary for most fluorescent gels/blots.

Chemiluminescence and Bioluminescence

Most luminescence methods yield much dimmer specimens than fluorescence. These dim specimens need to be imaged using film, or with a highly sensitive imaging apparatus. Film images of luminescent gels/blots can be analyzed by any of the methods described above (see Brightfield Illumination). Multipurpose imaging plate readers can also be used for some luminescence assays. For a discussion, see the Imaging Plate Readers booklet.

Cameras

The brightest forms of chemiluminescence can be seen with a moderately sensitive integrating video camera. However, many chemiluminescent and most bioluminescent specimens, are quite dim, so a high performance camera is required. We recommend a cooled CCD camera capable of digitizing at 12-16 bit precision. This type of camera will be able to image the majority of specimens.

With a moderately sensitive cooled CCD camera (such as the Hamamatsu C4880), we obtain usable images of typical chemiluminescent Southern blots in about the same exposure time as is used for film. If you need a 3 minute film exposure, you will need about a 2-3 minute exposure on a good cooled camera. This becomes tedious with fainter specimens (e.g. 10 minute film exposure). While any number of film specimens can be exposed simultaneously, a camera images just one specimen at a time. If that imaging exposure takes 10 minutes, it can really limit throughput. There are also problems in timing the imaging of flash luminescence (e.g. ECL). If the image acquisition is not just right on the first try, the specimen may have faded by the second attempt. It would be advantageous to speed up the imaging process.

The best way to speed up the imaging process is with a highly sensitive camera. Some CCD detectors approach 80% quantum efficiency, and one of these with a high precision digitizer (16-bit) will be faster than a standard cooled camera. We can supply a gel/blot analyzer incorporating this type of state-of-the-art camera. It will make fine images, but it is costly. An alternative is our Midnight Sun camera, which is fast, relatively low in cost, and needs no liquid cooling. With Midnight Sun, specimens requiring 3 minute film exposures can be imaged in less than a minute. Both flash and glow luminescence (almost any specimen emitting in the 400-850 nm range) can be imaged. The major limitation of Midnight Sun is a dynamic range of about 50:1. Bright bands will tend to flare if you try to image faint bands in the same exposure.

Calibration
Accurate calibration in both the density and spatial dimensions is essential for many applications. MCID and AIS are fully open to calibratation, and include the following:

Artifact Correction

Gels/blots contain artifacts, including variable density baselines that affect densitometry, and warps that affect molecular weight measurements. MCID and AIS include functions designed to minimize artifacts.

For example, most gels/blots contain a baseline component. MCID and AIS remove baseline, either by processing the image to isolate relevant bands or spots, or by processing the numerical data extracted from the image. In both cases, baseline correction is fast, automated, and yields results that correspond well with baseline corrections applied by a skilled technologist. Editing of the machine-defined baseline is rarely necessary, but is easily performed.

Correction of spatial artifacts is also available. MCID and AIS include "unwarping" functions, that can remove artifacts such as stretching or smiling. The corrections are applied easily and rapidly enough to be used routinely.

Archiving and Documentation

Gels are fragile, and tend to deteriorate with handling and with time. Therefore, laboratories use photo documentation systems to create permanent records of the specimens. Photo documentation is relatively costly (about $1/print) and tedious to use (often multiple exposures are necessary to obtain a good picture). Further, there is very little that can be done with a photograph. It cannot be transferred to computer applications, such as word processors. It cannot be sent to other laboratories via computer. The photograph can be analyzed, but quantitative analysis of photographs is subject to errors caused by nonlinear film and paper response, and reflections. It would be much better to analyze the gel, directly.

When the limitations of photo documentation become bothersome, laboratories evaluate digital archiving systems. In a digital archiving system, some form of camera or scanner is connected to a digitizing system, and is used to create a digital image of the gel. The digital image is saved as a file on a computer disk, and serves as a permanent record of the gel. The digital image can be retrieved from the computer and displayed at any time, and it can be printed using many photographic or video printing techniques. However, low cost digital archiving systems (cost ranges from about $7,000 - $15,000) are inflexible and include only primitive image analysis software. Often, the file containing the digital image must be transferred to other systems for quantitative analysis.

Advantages of an Integrated Archiving and Analysis System

A digital archiving system should be selected with attention to future needs. If quantitative analyses will be required, now or in the future, consider installing a single system that can handle both the archiving/documentation and analysis of digital images. It will be a bit more complex and costly, initially, but will save both time and money in the long run. Some advantages of an integrated system are summarized, below.

Use of calibrated images

Digital archiving systems produce image files in various desktop publishing formats (e.g. TIFF). It is these files that are transferred to image analysis software. A TIFF image contains no density or spatial calibrations, and this leads to a number of difficulties:

• the image analyzer must be recalibrated each time the gel image is retrieved for analysis;
• calibrated images cannot be exchanged between image analysis sites;
• there is no easy way of demonstrating an integral image calibration for the purposes of quality control or regulatory compliance.

In contrast, MCID and AIS images can be saved in TIFF format, or in MCID's own calibrated format. Because MCID images include full calibrations, the images can be sent to any other MCID or AIS system, and can be analyzed there using the included calibrations. The images can also be retrieved at any date after an analysis, and the calibrations that were used to acquire each data point in the analysis report can be demonstrated. The saving of calibrated images adds greatly to the convenience of image archiving, and simplifies your data reporting procedures.

Only one system needs to be learned and validated

Because MCID and AIS include a full set of acquisition, archiving and analytical functions, only one program need be learned to do everything your lab needs. Also, it is only necessary to validate the quantitative performance of a single, integrated system.

Various types of image acquisition devices can be used

Digital image quality is limited by the primary acquisition device, and proper selection of acquisition components (see the guidelines above) is critical. Our image analyzers support a variety of primary acquisition devices.

Many forms of gel documentation are available

Archiving and printing of the gel/blot image is only one aspect of the documentation task. Many laboratories also need prints and slides containing graphs and numerical data. MCID and AIS handle all aspects of gel/blot documentation. They create data reports, graphs, and tables, and export data to popular spread sheet programs. For image hard copy, they support a broad variety of printers, and digital and analog film recorders.

For many applications, absolute density precision is not critical, and the major selection criterion is a reasonably even field of illumination at the required wavelength. A good basic transilluminator will offer relatively even white light or UV illumination. This type of illuminator is available from many sources. If precision white light transillumination is necessary, we recommend our Northern Light. This instrument provides cool, homogeneous light, feedback regulated to within ± 0.5% over a 12 hour period. Homogeneity of illumination is also superior, as we use dual concentric circular light sources, rather than the straight tubes used in other illuminators. Finally, light intensity is continuously variable over a broad range, with power to the lamps displayed on a digital readout.

Transilluminators are best suited to gels run in a transparent matrix. However, many specimens have been transferred to non-transparent membranes, are run in a non-transparent matrix, or are analyzed from printed photographs. Dorsal illumination systems are used for working with these types of non-transparent materials. It is common for both transillumination and dorsal illumination to be available within a single illuminator. For example, some devices offer white light or UV transillumination, and dorsal illumination that may be selected from white light, short, and long wavelength UV.

Brightly lit specimens (e.g. film autorads) can be imaged in a standard room environment without worry about contamination of the image by ambient illumination. Fluorescent and luminescent gels/blots, in contrast, need to be isolated from ambient illumination. An isolation chamber can be a simple curtained box for UV gels. For direct imaging of chemiluminescence, it will be necessary to use exposure conditions that are equivalent to those in a photographic dark room. Any light leaks in the exposure chamber will swamp the very low levels of emitted light.

Filtering UV Illumination

The light emitted by UV gels can be dim. In addition, fluorescence from the specimen is mixed with non-specific illumination from the lamps. To yield the most sensitive analyses, we must remove as much as possible of the non-specific signal. At minimum, the specimen is viewed through a high pass filter that removes any wavelengths below about 400 nm. This removes the contribution of the UV illumination, and leaves the emitted fluorescence as the major source of signal. Even better, we could use a band pass filter that is tuned specifically to the wavelengths at which our stain is emitting relevant information. For example, ethidium bromide is excited most efficiently at UV-B wavelengths (about 312 nm). It emits fluorescence extending from about 500 - 700 nm, peaking at 590 nm. We view the specimen through a 580 - 640 nm bandpass filter. This removes irrelevant contributions from the illuminator and from ambient lighting, while passing the emission from the stain.

Duewer, D.L., Currie, L.A., Reeder, D.J., Leigh, S.D., Liu, H.-K. and Mudd, J.L. Interlaboratory comparison of autoradiographic DNA profiling measurements. 2. Measurement uncertainty and its propagation, Analytical Chemistry 67:1220-1231 (1995).

Mansfield, E.S., Worley, J.M., McKenzie, S.E., Surrey, S., Rappaport, E. and Fortina, P. Nucleic acid detection using non-radioactive labeling methods, Molecular and Cellular Probes 9:145-156 (1995).

Patton, W.F. Biologist's perspective on analytical imaging systems as applied to protein gel eletrophoresis, Journal of Chromatography A 698:55-87 (1995).

Ramm, P. Advanced image analysis systems in cell, molecular and neurobiology applications, Journal of Neuroscience Methods 54:131-149 (1994).

Reichert, W.L., Stein, J.E., French, B., Goodwin, P. and Varanasi, U. Storage phosphor imaging technique for detection and quantitation of DNA adducts measured by the 32P-postlabeling assay, Carcinogenesis 13:1475-1479 (1992).