Scanner Settings

Scientific scanners and imagers are complex instruments. The following section should give you some background information to make the right choices for scanning your gels.

The principle of operation of a scanner is quite simple: a pixel in the image file is created from light intensity that is measured at a certain point in the gel. In more detail, light is emitted from a light source, reflected or transmitted by the gel, and then measured by a detector. The measured value is converted to a number by an A/D (analog to digital) converter. The numbers from the whole gel are then processed to give the image file that is the output of the scanning process.

Choosing Scan Mode

The Scan Mode describes how the light travels from the light source to the detector.

• In transmisson mode the light passes through the gel and is directly measured by the detector on the opposite side of the gel (see left figure above).
• In reflection mode the gel is between the light source and a reflector (normally in the scanners lid). Light travels through the gel, is reflected, goes again through the gel, and is then detected.

Three scanning principles: light transmission, light reflection, fluorescence

For high quality data we recommend using transmission mode because saturation effects are less likely than with reflection mode. Saturation means that differences in higher spot intensities are not properly resolved -- dark spot centers become completely black. In reflection mode this happens more often because the light travels twice through the gel, so it can be more easily absorbed completely. Saturated regions appear as "plateau spots" in the 3D view of a spot (see the following figure).

Saturated "plateau spot"

Consideration of the scan mode is important if you use staining techniques absorbing the light, e.g. all types of Coomassie or silver nitrate staining. For fluorescence scanning the scan mode is less relevant because the light is emitted by the proteins within the gel.

Light intensity

The intensity of the light source directly influences differentiation of high density spots. It is common to use office document scanners to detect absorbing dyes (Coomassie, Silver, x-ray film). The manufacturers usually claim a differentiation of gray levels until an optical density (OD) of 2.0. This is mostly sufficient for gel documentation purposes. However, it is insufficient for differentiation of high density spots reaching optical densities near 3.7 to 4.0.

Document scanners resolving ODs from 3.7 to 4.0 can be found in the premium or professional product lines of well known manufacturers (e.g. Quato, Umax).

Warming up the lamp before starting the scan process is highly important because the lamps become more powerful when they reach their working temperature. Extensive usage of the scanner causes aging of the light source resulting in weaker light intensity.

Especially for fluorescence imaging high performance light sources are absolutely necessary. In this field of imaging normally lasers or UV lamps are used. Lasers deliver highly concentrated light and can effectively excite fluorescent dyes. Depending on the fluorescent dyes different colored lasers and excitation and emission filters are needed. Lasers need a warm up phase too to reach their highest performance.

The light harvesting detector and the A/D converter collect the light that has passed through the gel and translate this information to computer readable signals. The detector (CCD element or PMT) converts the light intensity to an electrical signal that is subsequently digitized by the A/D (analog to digital) converter. These data are stored in the image file.

In many high perfomance imaging devices for fluorescence detection also very weak signals have to be detected. That is why these scanners allow to tune the voltage of the light detector (PMT - photo multiplier tube). Higher voltages increase the sensitivity for very weak signals but also increase image background. Unfortunately increasing PMT voltage also results in amplifying noise. Because of this fact PMT voltage should be modified with care. Another pitfall is simultanous enhancement of strong signals bringing very intense spots to saturation.

In desktop scanners the A/D converter very often can differentiate a larger dynamic range of intensities than what can be stored in the file format. This is the reason, why the internal (A/D converter) color depth (e.g. 12 bit) may be larger than the output (e.g. 8 bit). The advantage of such a constellation is the flexibility for the selection of the right intensity range after a prescan for the final data storage. An alternative approach is the data processing of the complete data coming from the A/D converter by using gradation curves (see also image calibration).

Intensity range

Enlarging the intensity range between signal and background is one of the major tasks in image generation. The larger the distance between background and top intensity on an image the larger the dynamic range that is available for coding intensity levels. The more intensity levels are coded in the image the more accurate is the quantitative analysis. There are several ways to increase the signal to background range:

• Cutting off a high background by a scanner preset. In the prescan the user can traditionally specify the lowest and highest intensity signals that will be used as lowest and highest intensity signal in the resulting image file. The scanner will then distribute the measured intensity values evenly between the given minimum and maximum. This makes sense if the A/D converter can distinguish more intensity levels than what can be stored in the image file. Note that this approach only reaches highest performance if you avoid scanning areas that are not covered by the gel.
• Controlling the conversion from color to grayscale. Most office scanners create grayscale images by converting a color scan to a grayscale image file. This is done by averaging intensities over the different color channels. You can improve the resulting image in some situations by controlling the color to grayscale conversion yourself. For example:
  • A coomassie blue stained image can be better scanned by excluding the blue color channel from the color image befor converting it to a grayscale image.
  • A similar approach can be used for the brownish silver stains. As a rule, extract those color channels with high absorption. Brownish silver stains absorb the blue light most efficiently, so extracting the blue color channel will give the best results.
• Increasing the voltage of the light detector. If the intensity range is not completely used by intensity signals in a fluorescence scan the voltage of the light detector may be increased. This will increase the signals overall but also increases the absolute distance between background signal and strongest intensity signal on the gel image. Please be aware of saturation effects!

A perfect spot without any saturation effects

Saturation effects heavily affect quantitative data because they virtually truncate the peaks of high intensity spots. These truncations distort the quantities of spots. Many scanner software packages highlight saturated image regions. Saturation effects may occur for several reasons:

Intense staining can cause complete absorbtion of light passing through the spot. This results in planar truncated spots in the 3D view of the gel image. Try to use a scanner with a more intense light source (e.g. a laser scanner) or to warm up the scanner's lamp for a longer time. Reducing staining or increasing destaining time may also solve this problem.

Dyes can appear in high concentration. Some staining techniques accumulate pigment on top of high intensity spots losing the amount / intensity correlation. This can be observed in all techniques based on the photographic silver nitrate staining (silver staining, x-ray film).

While for classical silver staining techniques the typical doughnut spots may be observed the saturations from x-ray films have a planar appearance. To avoid these problems use shorter staining or developing times, or load less protein on the gel. Silver staining should not be used in quantitative proteome analysis because of saturation problems, and, more importantly, because the relation between silver stain intensity and protein amount is usually not linear.

A saturated protein spot

A similar effect can be observed during phosphorimaging. If a gel is overexposed to a phosphorscreen the screen's capacity to accumulate energy from radioactive decay events may be exhausted. This also results in planar truncated spots in the 3D view. Here reducing the exposure time will solve the problem.

Heavily fluorescing gels may overload the light detector or heavily excited light detectors may overload the A/D converter. This results in a loss of differentiation ability of strong signals. This also results in planar truncated spots. Here reduction of the PMT voltage may help.

Silver stain: A "doughnut protein spot" that lost it's central quantity.

Subsequent manipulation of the image data by using image processing software not designed for quantitative image analysis may ruin your images. Visually enhanced images often show truncations in the high and low intensity regions. Gel image files are not only pictures but also collections of quantitative data. Only flipping, morroring, rotating in 90 degrees steps and cropping leave the original data intact. Free rotating, enlarging or reducing, any kind of gray level, contrast or gamma adjustion distorts the quantitative data and may destroy your image files (see also image calibration).