3.65 MB

| Deutsch | | Español |

Warping with Smart Vectors^TM HQ Technology

Background: Why warping is important for 2D gel image analysis?

Usually, spots from the same protein appear at slightly different positions on different gels. These positional differences are the main reason for inefficiency in 'traditional' 2D gel image analysis packages.

Gel image with overlaid grid	The same gel image after warping

If you are used to classical 2D gel analysis packages, you might know those long and te-dious editing sessions for spot matches and spot boundaries that were found "automatically" by the software.
Delta2D's image warping eliminates these running differences between 2D gel images before doing spot detection or quantitative analysis using an image processing method called Warping. The result is like having perfect gels: after warping, corresponding spots have the same position.
DECODON pioneered the use of image warping in 2D gel analysis with the release of Delta2D version 1.0 in 2000.

DIGE and other multiplex methods allow you to co-migrate multiple samples on the same gel. Thereby they mitigate the problem of differing spot positions. However, as soon as an experiment contains more than one gel, gel analysis software will have to deal with different migration positions.

Here you see a combined image that is made from two gels. One is colored in orange, the other one in blue. In the dual channel image without warping it is hard to find corresponding spots and do comparisons of expression patterns.

Two gel images, overlaid into a dual channel image, not warped.	Two gel images, overlaid into a dual channel image, after warping. Differences in expression levels are clearly visible. Warping allows Delta2D to assign corresponding image positions across a whole set of images. This enables, for example, 100% Spot Matching and Image Fusion.

After the warping, the dual channel image gives valuable insight for comparing the spot patterns qualitatively. Blue spots are stronger on gel A, orange ones on gel B. Dark colors mean that spots have roughly the same intensity on both images.
But the advantages of warping go far beyond the making of dual channel images. The effect of applying image warping is as if you had made perfect 2D gels: those would have all proteins migrating exactly to the same position. And because Delta2D knows about pixel-by-pixel correspondences between images, a number of Delta2D's other core technologies are now available:

• 100% Spot Matching
  Delivers you higher statistical confidence for the analysis of expression profiles.
• Image Fusion
  Lets you combine multiple images to produce, for example, average images.
• Proteome Maps
  Use union fusion images e.g. to store protein identifications.
• Color Coding
  Analysis of Expression profiles with spot color coding combines image fusion and 100% spot matching to deliver a high-level overview of groups of proteins having similar profiles.

Defining a warping strategy

The warping for a whole experiment in Delta2D is based on pairwise warpings between gels. You do not have to produce match vectors for every possible gel image pair in your project: Given a warping between gel A and gel B, as well as a warping between gels B and C, Delta2D will automatically create an implicit warping between A and C.

Implicit gel image warping

A 'Warping Strategy' contains the definition of pair-wise warpings that Delta2D should use to construct the overall warping. It can be flat or hierarchical.
Choose 'Project -> Warp Strategy' from the menu of the Project Manager, select the Group Warping Strategy, and define the Automatic warping for warpings within as well as between groups. Then press OK to apply the selected strategy.
With the Group Warping Strategy you warp the two gels within every replicate group using the automatic warp mode. Connect the replicate groups by warping the first gel of the 'control' group (control_01) to the first gels of the '1min' and '10min' groups, again using the automatic warp mode.
The Group Warping Strategy is suitable for most projects with replicate groups. The general guideline for setting up a warping strategy is to connect similar gels directly, and to make sure that every possible gel pair can be warped by combining a few direct warpings. See the manual section 'Defining a Warp Strategy for the Complete Project' for more information.

Warping Strategy Manager

You could also define the warp mode for every single cell in the project manager using the context menu. Using the Warping Strategy dialog is, however, easier and less error-prone.

Automatic warping employs the SmartVectors^TM HQ technology and delivers the result as a set of match vectors that connect corresponding spots. Always check the resulting dual channel image and if necessary improve the warping iteratively as described later in this document.

The Strategy Manager overrides previous manual assignments of warping methods. So apply it after having organized a new project and be careful when applying it again later. Existing sets of match vectors are not affected.

In the Project Manager window, you see now that the display of the cells representing gel image pairs has changed:

Setting the warp strategy. Notice the tool tips that appear when moving the mouse pointer over the cells or the images.

As explained above, the warping strategy is to use automatic warps to connect images within the same group, and to connect the first image in every group to control_01. The Project Manager reflects this: a yellow circle means that the images in that row and column should be warped directly. All other possible gel image pairs can be warped by combining one or more of these warpings.

As a guideline, you can think of the yellow circles as showing you where you have to do work for the warping step. By creating match vectors for each of these image pairs, we allow Delta2D to warp the whole project.

When using the Automatic Warping you can let Delta2D run these warping jobs Automatically. Choose 'Window -> Job Manager' and click the 'Run' button. The Job Manager will then monitor and perform automatic warping jobs.

Setting up a warping strategy for DIGE
and other multiplex projects

Delta2D offers a special warping strategy for DIGE and other multiplex experiments that use an internal standard. Within the same gel, we will not have to do any warping because samples are co-migrated. For warping between gels, we use the internal standard images because they always show the same internal standard sample which makes them very easy to warp.

Before you start, make sure that you have checked 'use internal standard' in project properties, and assigned gels and internal standard images.

Choose 'Project -> Warp Strategy' from the menu of the Project Manager, select the 'In-Gel Standard Warping Strategy'.
Warp mode within gels should be set to Identical, Warp mode between gels should be set to 'Automatic'. Then press OK to apply the selected strategy.

Warp mode 'Identical' means that Delta2D will take the images as-is, without doing any warping. This is suitable for images from the same gel. Of course, you can change this setting if you want to apply Delta2D's warping (e.g. if there is a minor shift between the different images).

Warp Manager showing In-Gel Standard Warping Strategy

SmartVectors^TM HQ: Find Match Vectors Automatically

The goal of this step is to produce a warping that aligns spots on one image with corresponding spots on the other image. For every direct warping, you see a yellow circle in the project manager. Warping is based on so-called match vectors. Delta2D's SmartVectors^TM HQ Technology assists you in producing these match vectors quickly and reliably.

In the Project Manager you see a tabular arrangement of all images in the project. In order to open a gel image pair in the dual view, point the mouse at the table cell where the row of one image (e.g. 'control_01') meets the column of the other (e.g. 'control_02'). Then press the 'View' button.

Dual Channel image of two unwarped gel images

A new window (the dual view of these images) opens and shows a combination of the two images (e.g. having colored 'control_01' in blue and 'control_02' in orange). Thus, you can see the differences in spot positions at a glance: black or gray spots have about equal volume in both images. Orange spots are stronger on the image 'control_02' while blue spots are weaker. The more their color tends to one extreme, the bigger the difference is in expression.

If you are analyzing a DIGE experiment, please open the internal standard images from two different gels in the dual view.

Of course you can navigate any time through the dual channel image, zoom in and out or view the single images only. You can also correct the presentation of the images by choosing 'Histograms...' () from the menu 'Images' and by using the sliders to adjust the settings and clicking on the Equalize Image Button () in the tool bar to make sure that both images are balanced. Changes will affect the visualisation only, while quantification will always be done on the raw images. Please also check if the gel image background has been hidden by using the background button () in the tool bar. This adjustments help to better see the differences in spot positions. Confirm with OK.

Histogram settings dialog

Press 'Find Match Vectors' in the dual view now. Delta2D will use the SmartVectors^TM HQ technology to find a set of match vectors (the match map) and switch the warp mode to exact.

Delta2D distinguishes between match vectors that were set by hand (shown as solid lines), and those that were created automatically (shown as dashed lines). By 'approving' a match vector you declare that it was checked by you, so it can be used to guide the finding of more match vectors.

Now press the warp button () and Delta2D will warp the orange image in accordance to the match map using the exact warp mode. In the resulting warped view, it is usually easy to see if all corresponding spots are aligned. Regions that are not sufficiently aligned show up as similar spot patterns in the blue and in the orange part of the image. Delta2D will only create match vectors in regions where it is reasonably certain that they correspond. You can unwarp the images (press unwarp ) whenever you want to review the match vectors with respect to the original images.

If there are regions that are not properly warped yet, you can use the 'Find Match Vectors' button again. Delta2D's SmartVectors^TM HQ technology will use the vectors that are already present to guide the process of finding new ones.

Press the 'Find Match Vectors' button. Delta2D will ask you what to do with the vectors that were automatically created in the previous step: you can decide to use them as basis of another iteration with SmartVectors^TM technology or to reject them.

As soon as you have approved any match vector in the gel image pair, the symbol in the project manager will change from yellow to green, meaning that you have already worked with the match vectors. It is good practice to approve all match vectors when you are finished with a gel pair (use Matches -> Select Non-Approved, and then Matches -> Approve Selected).

Creating, deleting, and changing match vectors by hand

In order to work with match vectors, you have to select the Match Vector Tool first. Click on the top-most button in the vertical tool panel at the left of the dual view window.

Sometimes you will want to correct or improve the match vectors that were found auto-matically by Delta2D. The general approach is to work with whole image regions (as op-posed to single spots). If one or more match vectors in a region are not correct, you just delete all match vectors in that region and press 'Find Match Vectors' again to find better ones; vectors from surrounding regions will guide the process.

The Dual View after Warping

Of course, you can set and change all match vectors by hand if you want:

• Selecting a single match vector
  You can select a single match vector by clicking on it. Selected match vectors are highlighted.
• Selecting match vectors in a region
  Drag a rectangle with the mouse (keep the left mouse button pressed). All vectors in that rectangle will be selected.
• Deleting match vectors
  Right click on a match vector and choose 'Delete' or 'Delete Selected' from the menu. This will delete the selected or all selected match vectors, respectively.
• Setting a new match vector
  Click first on the spot in the orange image, then on the corresponding spot on the blue image. Hold down the CTRL key while you click in order to let automatically snap the ends of match vectors to spot centers.
• Changing a match vector
  Drag one end of the match vector in order to change it. Hold down the CTRL key while you drag in order to switch off spot snapping.

You can undo match vector operations by selecting 'Undo' () from the Edit menu.

Click on the Warp button to apply your new match vectors. You can iteratively add, correct or delete match vectors until the result is satisfying.

Verifying the results

Since the quality of the whole analysis relies on the exactness of warping, it is worth working carefully in this step of analysis. In order to control the quality of the warping step, please check this list:

• In the project manager, you should see cells with green circles only, or with no circle.
• For each gel pair (cell) with a green circle, open the dual view and check the warped dual channel image.

Even if you set the warp mode to automatic you should still review the resulting warping in the Dual View to check if you are satisfied or if a correction is necessary.

The demo data set includes prepared match maps. To save time, just import the respective match map for a pair of gel images in the Dual View. In the menu choose Matches -> Import. If the blue match vector ends point into orange spots and vice versa the match vectors have the wrong direction and you need to convert them by choosing 'Matches > Invert'.