Chapter 5
Warping 2D Gel Images

5.1 The Gel Image Pair View

The Gel Image Pair View shows a gel pair and lets you create or refine a warp transform between them.

Figure 5.1: The Gel Image Pair View, also named Dual Image View, or, in short, Dual View

The images are loaded after a fixed scheme: the gel image in the table row is loaded as the master image, the other one (in the table column) takes the role of the sample image that will be warped.

From the gel images, a dual channel image is automatically generated: the master image is colored blue while the sample image will be displayed in orange color. Shades of black¹ are generated where both images have dark regions, i.e. spots. You can click on the tabs at the bottom of the Dual View window to switch quickly between displaying the master, sample, or dual channel image. You can use as well Ctrl + Space and Ctrl + Shift + Space to switch between these tabs.

________________________________________________________________________________

Note:

Delta2D can use any color scheme to produce a dual channel image. When you first start it, it will be set to the default color scheme: white for background, a shade of blue for master spots, and a shade of orange for sample spots. Regions with overlapping spots are colored black. For changing the color scheme, please refer to section 5.13.

Figure 5.2: The tabs for controlling image visibility

The Tool Panel

Figure 5.3: The tool panel.

Match vector tool. In match vector mode, only the match vector overlay is visible, mouse clicks define corresponding gel positions. Spots tool. Only the spots overlay will be visible. If no spot boundaries are present, spot detection and quantitation will be started. On present spots, this tool enables you to select, mark, hide, and cancel spots, and to exclude spots from or include them into normalization. allow selection of spots and spot groups. Edit Spots tool. Click here to switch to the edit spots mode. Available edit markers become visible, you can add, split and fuse spots (details in sec. 6.5). Zoom tool. This is the initially active tool. All other layers (except edit markers) are visible. Use this tool to control image zoom by clicking into or dragging a rectangle around interesting regions of the image. Labels tool. The labels overlay remains or becomes visible, the other overlays are hidden. Use this tool to create and edit labels. Table 5.1: Buttons on the tool panel.

The Icon Bar

Figure 5.4: The icon bar of the Gel Pair View

Export the warped or unwarped sample image as image to disk. Export the dual channel image as image to disk. Open a snapshot window, that contains the whole gel image as it is now. Open the preferences dialog Zoom in Zoom out Zoom 1:1 Fit the image into the window, such that it can be seen completely inside the window. Import a match map Export the current match map Warp Apply the current warp mode to the sample image Disable warping operations and show sample image in its unwarped state. Current warp mode: Select the warp mode for this sample gel image. Find Match Vectors: Apply the SmartVectors Technology to receive an automatically genreated match map. Show image histograms Equalize images Show or hide the foreground of images Show or hide the background of images Help. . . About. . . Table 5.2: Buttons on the icon bar and what they do.

Note:

The icon bar, the tool panel and any similar looking component of Delta2D can be torn off and placed anywhere on your screen by clicking on its "handle" at its beginning and dragging it to the desired place. To reattach it to the main window, simply close the small window of the respective bar.

The Status Bar

Figure 5.5: The status bar of the Gel Pair View

Position and zoom level

The left field of the status bar shows the coordinates of the mouse cursor in relation to the gel image. This makes it easy to evaluate the mouse position in a higher zoom level, to locate a certain spot quickly, as well as to access the size of a spot. In brackets near the coordinates, the zoom level is indicated.

Image Size, Match Vector count, Spots count

The middle field indicates the outer bounds of both images. This means the horizontal size of the broader of both images and the vertical size of the higher one. This value refers to the original images, no matter how the prescale is used. Besides, you get a quick information on how many match vectors and how many spot boundaries are present in this pair of images.

Memory use

The right part shows the amount of memory currently used in relation to currently reserved memory. Double click on this field to free unused memory.

5.2 Configuring the Display Using Rollups

Rollups are little windows that are floating above the main window. They can be collapsed to use only a minimum of screen space.

The Colors Rollup

Move the mouse pointer over the dual channel image and watch the color square. A small circle inside the color square shows the current location in the color diagram. A numerical display of the intensity ratio (sample / master) is shown below the color square. Even though these values are computed only for small regions of the gel images, they are already quite useful indicators for the expression ratio of a spot.

Click inside the color square to change the color scheme. A menu appears, giving you the opportunity to change the used color scheme (see sec.5.13) and to switch directly between the absolute mode (standard setting) and the ratio mode (see sec.5.13).

Figure 5.6: The colors rollup

The Overlays Rollup

Figure 5.7: The overlays rollup

Clicking one of the small control buttons will toggle the visibility of the respective objects between three modes: visible, non-visible, and auto-visibility (controlled by context).

The button Gel images only is useful to blind out the overlaid objects temporarily for a short, undisturbed glance on the original images, e.g. during spot editing. The overlays will reappear according to the overlay rollup settings when you release the button.

The Navigator Rollup

Figure 5.9: The navigator rollup

The Zoom Rollup

Figure 5.10: The zoom rollup

The Expression Profile Rollup

Figure 5.11: The Expression Profile rollup. The columns are colored according to the replicate group. The black lines indicate the mean plus minus the relative standard deviation.

The 3D Spots Rollup

• show single spots, whereas the shown region dynamically adapts to the size of the chosen spot (default setting), or show bigger regions of a fixed size
• show the spots opaque or as wire frame model (useful to make interlocking spots visible when viewing both gels in the Dual View)
• change the color of spots and background
• adapt the height scale to your needs (e.g. for very flat or very tall spots)

Choose Preferences... from the File menu of the Dual View and switch to the 3D Spots tab to change the settings. Please refer to section 9.10 for more details.

Figure 5.12: The 3D spots rollup

Using the 3D Spots Rollup

• click with the left mouse button and drag the area within the rollup, you can turn the view in any direction
• To zoom in or out, drag up and down while holding the middle button pressed
• With the right mouse button you can move the scene inside the rollup in any direction without changing the view angle
• When wiewing an area with multiple spots just click on a spot inside the rollup and watch the same spot being selected in the Dual View. Thus you can e.g. regain orientation on a 3D view in case you lost it, without shifting the viewed area , or simply locate interesting spots.

The pI/MW Calibration Rollup

The rollup shows the estimated pI and MW for the current mouse position, it updates as you move the mouse. Delta2D can estimate the isoelectric point and molecular weight of a spot according to known data for other spots on the gel. This is accomplished by building a mathematical model of the relation between pI/MW values and a spot's position on the gel. The known data is taken from label aspects (Scout data), in the pI and MW attributes. More spots with known pI/MW values make the model more accurate.

Figure 5.13: The pI/MW Calibration rollup

Configuring the Data Source for pI/MW estimation

The precondition to configure pI/MW estimation is the availability of applicable reference data in any scout (for more about scouts please refer to section 7.7) of some labels. This can be the manually edited physicochemical properties scout, the table data scout with arbitrary defineable data fields, or any web scout providing adequate data.

Go to the Project Manager and select Preferences from the File menu. Switch to the Labeling tab. The Source field lets you choose from all scouts containing adequate data. Select the scout you want to use for the estimation and specify in the fields below which data fields are to be interpreted as pI and MW values.

Example You know the pI/MW values for 4 points on your gel and want to be able to see them for other interesting spots or regions. Here is how to proceed:

• If not already existing, place a label on every point you know the pI and MW of. The label can but needs not be connected to a spot on the gel. The names of the labels are of no importance for this purpose; they can be e.g. simple consecutive numbers.
• Right click on the first label and choose from the context menu Edit scout data  Physicochemical properties, a new dialog shows.
• Insert the values for pI and MW in the respective data fields.
• Close this dialog with OK.
• Repeat the last three steps for every label you know the pI and MW for.
• Now open the Preferences dialog (File  Preferences from the Dual View) and switch to the Labeling tab.
• On the bottom of the left side you see three drop-down boxes. Click on the first one and make sure Physiochemical properties is selected.
• Click on the second drop-down box and make sure Isoelectric point is selected.
• Click on the third box and make sure Molecular Weight is selected.
• Close this dialog with OK

Now you can open the pI/MW Calibration rollup and just read the pI/MW for any region on this gel you are pointing to.

Figure 5.14: Setting the data source for pI/MW-Calibration

The Rollups Menu

5.3 Navigating in Images

There are two possibilities to adjust the view: The zoom tool bar for quick access to predefined views and the zoom tool for precise determination of the current view. With the buttons and you zoom in (resp. out), whereas the button resets the view to the natural size of the image. With you fit the image size to the actual window size.

The zoom tool allows you to adjust more precisely the region you want to magnify. It is activated by using the tool panel in the top left area of the main window (Figure 5.3).

Press the zoom tool button to activate zoom mode. The mouse cursor will change to a magnifying glass. Click anywhere inside the image to enlarge it. Click and drag to specify a region that should be zoomed in. Clicking and dragging with the right mouse button will change the mouse cursor temporarily to act as a magnifying glass as long as the right mouse button is pressed. If you need a magnifying glass also while working with the label tool or others, it might be more convenient to use the Zoom Rollup (See Figure 5.10).

Use the Navigator Rollup (Figure 5.9) to shift the visible region of your image without changing your zoom ratio.

5.4 Enhancing the Images View

Controlling Background Display

Try it: Use Images Show background to switch background visibility on and off. You can use the tabs at the bottom of the Dual View window to switch between the gel images and the dual channel image. Often the dual channel image without background is clearer than the complete dual channel image.

Figure 5.15: A dual channel image with background switched on and off.

Figure 5.16: Visual background settings

Show visual background region in spot detection dialog

means that this setting will be handed over to the Spot detection dialog. It can be seen, when you open the Spot detection dialog, but will not be applied to quantitative data before you start a new quantitation.

Overwrite visual background region after spot detection

means the opposite way of communication: if you change this parameter in the spot detection dialog, it will be changed here as well.

The purpose of these options is to keep the visual background region in sync with the technical background region, thus the default setting for both options is checked. Please note that the visual background is linked to the background detection controlled by the quantitation dialog. With the default settings it has influence on spot quantity.

The Histograms Dialog

Adjusting histograms is advisable when your image does not use the full dynamic range (i.e. there is no bright white or no black in it), or if there is a homogeneous background.

Delta2D allows you to compensate for differences in brightness and contrast between master and sample gel images by adjusting the image histograms. This can result in production of clearer dual channel images. Histogram adjustments are controlled with the Histograms dialog, illustrated in figure 5.17.

Figure 5.17: The histograms dialog.

________________________________________________________________________________

Note:

Histogram adjustment is saved individually for each image. It affects the generation of dual channel images and the representation of images in the Gel Regions View, but leaves the spot quantities unchanged. It is meant to enhance the view of the gel images without changing quantitative data. This is in accordance with Delta2D's principle of leaving original data unchanged as far as possible.

Histograms or Amplitude Rescaling?

Delta2D offers two tools to correct the visual representation of gel images with poor contrast: amplitude rescaling and Histogram adjusting. They complement one another and can be used solely or in combination, just as required.

Amplitude rescale is a standard approach in image processing, which rescales the effectively used range of gray values in an image to the maximum usable range. For example: if you have an 8-bit image (which can define one of 256 gray values for each pixel, and let 0 being black and 256 being white), where the darkest pixel has the gray value 40 (dark gray) and the lightest 180 (light gray; quite a gray and shallow looking image), amplitude rescale represents the value of 40 as 0 (perfect black), 180 as 256 (perfect white) and the intermediate gray values rescaled respectively. Thus you have a much more plastic and vivid representation of your image without having altered the information contained. Depending on your gel images, the effect of amplitude rescale you can observe can vary in a wide range: the enhancement can be either hardly visible if your images already use a wide range of gray values, or it can even make spots visible you could not see before if your images use only a small bandwidth of gray values.

But of course this approach has a drawback as well: loading and displaying images in the Gel Image View takes slightly more time, especially if according to your zoom settings only a part of the image would be loaded, because for amplitude rescale the complete image has to be analyzed. Thus, the recommended setting depends on your gel images: if the effect of amplitude rescale is hardly visible, you can switch it off (please refer to section 5.5 on how to do so) to load images faster; if your images are generally more flat and poor in contrast, amplitude rescale can be a great help.

Amplitude rescale is no more than a rough pre-enhancement, which can be sufficient in many cases, but cannot take into account problems with artificial signals like speckles or gel breaks. This evokes the necessity of additional fine tuning the enhancement process. Histogram adjusting gives you the full control of the enhancement process. You can change the rescale settings smoothly while watching how the representation alters with your changes. But you can also let Delta2D adjust the histogram settings of an image automatically in consideration of another images contrast situation to make them both better comparable.

Automatic Histogram Equalization

Manual Histogram Equalization

The current gray scale histogram of both images of the Gel Image View is displayed in the dialog. The histogram display shows you how many pixels are contained at each gray scale level in the corresponding images.

To achieve optimal equalization please proceed as follows:

• Move the sliders of each image until the contrast settings fullfill your needs.
• Press the Equalize button and look at the vertical slider. If the vertical slider indicates that one histogram is higher weighted compared to the other move the left slider of that histogram a little to the right and press the equalize button again.
• Do this iteratively until the vertical slider is located exactly in the center.

If done that way you have perfectly equalized images in the dual view.

Manually Adjusting the Balance Between Two Images

Modifying the Histogram for a Single Image

Manipulating the slider positions allows you to suppress image regions containing pixel values outside a specified range from the display.

Managing the Histograms Dialog

Locking the dual channel image

Normally, if you make adjustments to the histogram of an individual image, the dynamic image display will change to show only that image. If you want to be able to watch the effect on the dual channel image while you make the adjustments, check the box marked Lock dual channel.

Resizing the dialog

Enlarging the dialog itself may give you better control over fine histogram adjustments, since the histograms and image display will be scaled up to fit in a larger dialog.

Repacking the dialog

After rescaling the dialog, you can use the keyboard shortcut Ctrl-* to tidy up the layout of the dialog.

Applying or discarding changes

Clicking the dialog's OK button will apply the changes to the images displayed in Delta2D's main frame. Clicking the dialog's Reset button will reset all the histogram values to the values that existed at the time the dialog was invoked. Clicking the Cancel button will dispose of the dialog without any changes being applied.

About the Histogram Adjustment Process

1. if the pixel is brighter than a given threshold, make it completely white
2. if the pixel is darker than another threshold, make it completely black
3. otherwise apply a linear change to the gray-level of the pixel

Adjusting histograms is advisable when your image does not use the full dynamic range (i.e. there is no bright white or no black in it) or if there is a homogeneous background.

You may also wish to keep the following points in mind when making adjustments to the image histograms:

• Most importantly, remember that adjustments to the image histograms affect only the visual presentation of your gel images – there is no affect on the spot detection and quantitation process.
• If you have the image background subtraction feature activated, histogram adjustments will be made to the images after the background has been subtracted.
• If you have enabled amplitude rescaling, both gel images will already utilize the available range of gray scale values. You only need to make changes to the histograms if the gray levels are different in the two images, or if you want to suppress lighter or darker regions of the images from the display.

5.5 Preferences for Image Preparation

The preferences dialog, reachable via File Preferences and there on the tab Image Preparation lets you set different groups of options as follows:

Figure 5.18: Preferences: Image preparation

Prescale

Scales larger images down to the given pixel count. Smaller images are not affected. Prescaling happens when images are loaded and is transparent to the user. Use this option to improve performance for large images. The scale factor for both images is determined by the first image loaded. The pixel count must be set to 50 at least.

Amplitude rescale

Enhance images by amplitude rescale when they are loaded. If this option is checked the gray values of an image are rescaled linearly so that the darkest pixel becomes black and the lightest pixel becomes white.

5.6 Visual Comparison of two Gel Images

Delta2D's approach to analyzing 2D gel electrophoresis images relies on advanced image processing technology that compensates for the differences in spot positions between gel images. These differences are due to variations in running conditions and the gel casting process. They are what makes comparing and analyzing 2D gel electrophoresis images so difficult and error prone.

When you hold two similar gel images next to each other, you may have the impression that the spot patterns on one gel are more or less a distorted version of the patterns on the other. The process of distorting (or "un-distorting") images is called warping. Delta2D's warping algorithms help you to generate dual channel images on which corresponding spots are perfectly overlaid. In the dual channel image, differences in protein expression levels can then be easily recognized. The same algorithm that is used in producing dual channel images will be used in the subsequent quantitation step to obtain accurate and reliable spot matching information.

Figure 5.19: A region of the dual channel image, before and after exact warp. Corresponding spots are overlaid exactly, allowing for easy identification of spots with changed expression level.

5.7 Warping Gel Images

In Delta2D 3.6 there is one warp mode per gel pair. It can be exact, global, automatic, implicit, or identical. Here is a more detailed description of the warp modes:

exact

All spots that are connected by a match vector will be perfectly overlaid after warping. Other spots will be warped according to match vectors in their neighborhood. This is the recommended warp mode in combination with the SmartVectors.

global

Corrects for global gel distortions such as growing or shrinking, rotation, stretching. Set a few match vectors then use global to see more correspondences. Global warp is a good start for setting more and more match vectors by hand. It is almost never suitable for producing the final dual channel image because there are local distortions, too.

automatic

Let Delta2D try to automatically find match vectors using the SmartVectors Technology and apply this match map to an exact warp. If match vectors are present they will be used to guide the automatic warping process. As with exact warping, spots that are connected by a match vector will be perfectly overlaid.

identical

This is just no warping at all. Identical transforms can be used for registering images that are from the same gel but display different samples or multiple aspects of a sample.

implicit

Warp the images by a combination of explicit transforms (these can be exact, global, automatic, or identical) that connects them. Example: Image B has a valid warping to image A, image C is also connected to image A, then image C can be compared to B using implicitly the existing warpings: C A B. You will usually have implicit warps for most of the gel pairs in your project.

The mode is used for producing the dual channel image as well as for matching spots in the quantitation tables.

Automatic Warping

Improving the Results of Automatic Warping

________________________________________________________________________________

Note:

You can employ the SmartVectors Technology and manually define match vectors iteratively to improve the result, if needed.

Read more at www.decodon.com/Support/Howto/SmartVectors.html

Do a Global Warp

Figure 5.20: A region of the dual channel image, before and after global warp. Corresponding spots are not overlaid exactly, but much closer than in the original image. Further correspondences can be identified more easily.

Do an Exact Warp

Add More Match Vectors if Necessary

Figure 5.21: An image region with well-aligned spots after exact warp (left image) in comparison to the same region after global warp (right image).

Note:

Whatever you have selected as the warp mode for the gel image pair will take precedence over the match vectors that you have provided. For example, when you have a match map and selected implicit warp mode, the gel images will be warped using whatever indirect connection is available, ignoring your match map. Switching the warp mode to exact or automatic will let Delta2D use the match map for warping.

5.8 More About Match Vectors

Match vectors connect corresponding spots (Figure 5.22). They are used by Delta2D to warp one image to another reference image to eliminate the differences in spot positions. The warping can be specified by using match vectors alone, or by using them to guide the SmartVectors Technlogy. Only a tiny fraction of all spot pairs has to be connected by match vectors because Delta2D uses a match vector to align an entire image region.

Before you start to set match vectors, make sure that match vector tool is activated by clicking in the tool panel.

________________________________________________________________________________

Note:

With version 3.4 we have introduced Undo and Redo for match vector operations. Thus you can try setting a couple of match vectors, study the resulting dual channel image and go back to previous match maps if you like.

Figure 5.22: Setting match vectors.

• Click on a spot in the sample image. It is marked by a solid circle.
• Click again to set the corresponding position in the master image. It is marked by a solid circle.

________________________________________________________________________________

Note:

It is important to draw all match vectors from sample (orange) to master image (blue).

Note:

In order to use automatic snapping of match vectors the option Match Vector Snap in the preferences dialog (Section 5.9) must be enabled.

Match Vector Snap

By default, any match vector you set will snap to the nearest spot maximum. This speeds up the setting of match vectors substantially. You can change this behavior using the Matching tab in File Preferences. To switch snapping temporarily, hold down the control key while you are creating or editing match vectors. Double clicking on the match vector will place both ends at their nearest spot maxima.

Creating Match Vectors by Double Click

Right-clicking on a match vector opens a context menu. Select Delete to delete the match vector, use Adjust to adjust both ends to their nearest spot maximum.

Saving the Match Map

________________________________________________________________________________

Note:

Saving match maps is not available in the evaluation version.

Saving the Warped Image

________________________________________________________________________________

Note:

Images exported in warped state can be used for documentation purposes only. They are not suitable for further quantitative analysis of any kind (reimported in Delta2D or imported in other software). Warping alters the complete image, which affects spot size as well. According to Delta2D's principle to leave original data unchanged, all quantitative analysis is done on the original, unwarped images.

5.9 Preferences for Match Vectors

Figure 5.23: Preferences: Match Vectors

Match vector snap to spot

If this option is selected, match vectors will snap to the nearest spot when they are moved around. This option can be temporarily toggled by pressing Ctrl while moving the match vector.

5.10 Defining a Warp Strategy for the Complete Project

To obtain satisfying results, it is worth to take one or two things into account:

• Since automatic works better the more the images resemble each other, it is most suitable for connecting replicate gel images, i.e. within groups. Gel images from different experiments are better warped using the exact mode.
• Warping one image on another always means that one image gets distorted, while the other (the Warping Master) remains undistorted. Since we use a control in most projects, it is very likely to use this as warping master and later on as basis for the proteome map.
• Avoid warping cycles as they can lead to unpredictable results. A warping cyrcle is a chain of warpings containing possibly contradicting directions: If you have four gel images A, B, C and D, there are warpings between A-B, B-D, A-C and C-D, and you want to set the warping mode A-D to implicit, Delta2D does not know which warping chain has priority: A-B-D or A-C-D.

Thanks to implicit warping we don't need to set up a warping between every single gel image. That's why this is the default warp mode when inserting images to a project. We only have to make sure that every gel image in our project is connected to at least one other image with a valid warping. There can be several intermediate steps (each of them exact, global, automatic, or identical) in-between two images. Again, this chain of intermediate steps will be used both for producing dual channel images and building expression profiles of the spots. Delta2D will always minimize the number of intermediate steps, and it will prefer manually specified warpings (exact, global, automatic with match vectors, or identical) over the implicit warpings.

Of course we could assign the warp modes manually between the respective images in our project. With four images this wouldn't be that hard, but later on, when having bigger projects, this can become quite a labor. Why waste time doing humdrum work if there is a more convenient way? Switch back to the Project Manager. Choose Project Warp Strategy . . . from the menu of the Project Manager or click on the button  to open the Warping Strategy Manager. This is simply a useful tool to automate the assignment of warp modes to the gel images of a project. It takes care that no gel is left out and no warping cycle is created accidentally.

Figure 5.24: Apply complete warping strategies at once

Note:

Please note, that the Strategy Manager alters assignments done manually before by setting every warp mode according to the chosen strategy. So use it at the very early stage of a new project, and don't touch it any more later on. Of course, warp modes can be changed manually at any time; then you have to take care for the consistency of your warping strategy yourself.

The Warping Strategy Manager (see figure 5.24) lets you choose between basic warping strategies:

 Group

This will be the most frequently applied strategy. It assumes that your image groups correspond gel replicates, and that the difference within groups is smaller than between groups. Within groups, images are warped with one warp mode (default: automatic) and the first image of each other group is warped to the first image of the first group with another warp mode (default: exact).

---

Figure 5.25: Group Warping Strategy

---

 Chain

All images of your project are connected like one long chain in the sequence of their appearance in the project, no matter to which group they belong. This strategy is recommended, if your samples have been taken at successive points of time in your experiment.

---

Figure 5.26: Chain Warping Strategy

---

 Chained Group

Combination of the two above strategies, applicable in case your image groups correspond gel replicates, and the groups represent successive points of time in an experiment.

---

Figure 5.27: Chained Group Warping Strategy

---

 One-to-all

Here one gel image takes the role of a master and all other gel images are connected only to this one.

---

Figure 5.28: One-To-All Warping Strategy

---

 In Gel Standard

Warp each standard image to the first standard image. Other images are warped to the standard image from the same gel, if possible. This is the default Warping Strategy for Projects using an internal standard and hence only available if the current project marked as DIGE Project. (Please refer to section 4.7 for more on DIGE Projects.)

---

Figure 5.29: In-Gel Standard Warping Strategy

---

5.11 Interpretation of the Dual Channel Image

Delta2D's dual channel imaging allows for quick and easy identification of changes in expression levels between gel images. Warping of the sample gel image onto the master gel image is essential for this: Since corresponding spots are now overlaid, the colors on the display give a direct indication of their relative expression levels (see Figure 5.30).

For better interpretation, you may wish to hide match vectors using the overlay rollup as described in section 5.2 (use Rollups Overlays to open the overlays rollup). Additionally, you may wish to switch off background or to adjust the histograms settings.

Figure 5.30: The result of the visual comparison process.

Move the mouse pointer over the dual channel image and watch the color square. A small circle inside the color square shows the current location in the color diagram. A numerical display of the intensity ratio (sample / master) is shown below the color square. Even though these values are computed only for small regions of the gel images, they are already quite useful indicators for the expression ratio of a spot.

For printing or presentation purposes, Delta2D's color scheme can be changed to an arbitrary combination of colors for master, sample, background and common pixels.

5.12 Questions and Answers About Warping

• How do i know which warping mode is the most adequate for my gel pair / project?
• By default, Delta2D uses implicit warp mode for any gel image pair. This is a reasonable choice because in the end most of the image pairs will be connected using implicit warps. Start by specifying warp modes for direct connections between image pairs in the Project Manager:
  • Between images from the same sample, choose automatic or exact warp mode. e.g. for a group consisting of replicates, connect each image of this group to its neighbor using exact or automatic warp mode, forming a chain within the group.
  • If you have multiple images per gel, choose identic warp mode between all images that were made from the same gel. If there were some experimental (except scanner reconfiguration and scanning) steps between recording of the images, try global, or, if this does not work, exact warp.
  • You will probably see some yellow arrows calling for your input. Add match vectors to each of these image pairs.
  • There will probably be some red interrupted arrows left, meaning that there is no chain of intermediate transforms connecting these images. Select one such pair with a red arrow and chose a direct warp mode. If the mode requires your input, you will recognize it by the yellow arrow. For choosing direct warp modes it is advisable to prefer gel images from the same run, or images in which you expect similar spot patterns. Another strategy is to connect the first image in a replicate group to the first image in the next replicate group.
  • If there appear red continuous arrows , you defined one explicit warping too much in the denoted pairs; a so called warping circle occurred. Reset one of the marked pairs to implicit warping.
  • You are ready when there are no more red or yellow arrows left.

  The project manager table should now show only green arrows, showing the direct connections. Thus, Delta2D has transformations connecting every image pair in your project and you can start analyzing expression profiles.

• I only want to do quantitative analysis, do I need to warp the images?
• Yes. And no. Delta2D uses the information from the warping to compute spot correspondences or an image fusion as basis for detecting a project wide spot consensus. That means you will have to create a match map (by hand or automatically) in order to see tables that show corresponding spots side by side. Furthermore, scatter plots require spot correspondences. In the Project Manager's table cell for a gel pair, you see what kind of warping information (manual, automatic, implicit) is available. But if you just want to obtain quantitative data for single gel images without comparing them to each other, of course, this is possible without match maps, too.
• Do I have to create a match map for every gel image pair in my project?
• No, Delta2D can combine match maps to connect two images via a third one, or even more indirectly using a chain of match maps. Thus, a good strategy is to connect all gel images in your project to a master gel image.
• What if one gel image is substantially rotated or shifted with respect to the other?
• Assign a few correspondences that you can identify reliably. Then use Warp Global Warp to eliminate global distortions. The global warping will bring similar spot patterns closer together, compensating for global distortions such as shifts, minor rotations, or differences in scaling. For a shift, already one single match vector is enough. If the rotation is > 90o, please rotate the respective image in the gel image properties dialogue (available from its context menu).
• What if initial match vectors are hard to find?
• With highly dissimilar gel images it is sometimes hard to find the first spot correspondences. Assign as many correspondences as you can identify reliably. Then use Warp Global Warp to eliminate global distortions, again bringing similar patterns closer together.
• When or why use global warp?
• Use the global warp early in the matching process, when you are not sure about further correspondences. Global warp is more robust with respect to wrong correspondences – one wrong match vector will not distort your image too much. Nevertheless, after the global warp, further correspondences will be easier to recognize.
• Does warping affect the quantitation process in any way?
• Not at all, spot detection and quantitation is done using the original images.

5.13 Using Color

Color Schemes

The Color Schemes Dialog

The Color Schemes dialog can be invoked from the Images menu, using the Images Color schemes. . . menu entry. Alternatively, you can click directly on the Color Schemes button in the tool bar.

Figure 5.32: The color schemes dialog.

Display Modes and Color Schemes

The currently selected color scheme is presented in the center of the Color Schemes dialog (figure 5.33).

Figure 5.33: The color schemes display.

• Top-left corner – The color used to display sample spots i.e. spots appearing exclusively in the sample gel.
• Top-right corner – The color used to display regions where spots overlap.
• Bottom-right corner – The color used to display master spots i.e. spots appearing exclusively on the master gel image.
• Bottom-left corner – The color used to display regions of image background.

If you are editing the scheme for ratio mode (see figure 5.34), you can choose two additional colors for highlighting points with only a relatively small ratio between the sample and master spot levels. Those colors are displayed in the middle of the top edge and of the right edge respectively. The ratio mode was implemented for a more fine grained representation of expression changes especially for faint spots.

Using Predefined Color Schemes

Figure 5.35: Some predefined color schemes

Defining Your Own Color Schemes

Create a New Scheme

Select New Colors for the Scheme

You can then choose a new color to use for highlighting the selected image feature. There are three methods available for doing this, accessible at the bottom of the Color Schemes dialog. The three color controls are described individually below.

Color swatches

This is the simplest color control to use. Simply select the color you want to use from the palette of available colors.

Hue-saturation-brightness (HSB) control

Using this tab, you can control the hue, saturation and brightness of a color separately. Select which of the three values you want to change by using the HSB radio buttons, and change the value by using the slider or by entering a value directly into the text field provided.

Red-green-blue (RGB) control

This control panel allows you to configure the levels of red, green and blue that are combined to produce the desired color. You can set these levels using the sliders provided, or by typing a value between 0 and 255 directly into the corresponding entry field.

It is recommended that you choose strongly contrasting colors for master spots, sample spots and the spot overlap, and an unobtrusive color for the image background, to help you to visualize the differences between two gel images quickly and clearly.

Renaming the New Scheme

Renaming a Color Scheme

You can't rename one of Delta2D's predefined color schemes.

Deleting a Color Scheme

You can't delete one of Delta2D's predefined color schemes.

Using Ratio Mode

Ratio mode works best with images that have a low background level. Therefore you should switch off background using the layer control and adjust the histograms, if necessary. Select Images Ratio mode to activate ratio mode display.

Figure 5.36: A gel region shown in ratio mode.