Dot Matrix Help

The dot matrix is a graphical representation of the scores for the complete pair alignment space (see below for an explanation of a dot matrix). One of the two sequences defines the X-axis (horizontal) and the other sequence defines the Y-axis (vertical). Every point on the graph represents the score at or near that point in the sequence comparison. By default, this is the local score, meaning that it represents the score quality only in the immediate region. You can also view the global (or cumulative) score, which shows the best alignment score up to that point in the alignment. In both cases, the darker the dot the higher the score. All zero and negative scores are white. Click the "-" and "+" buttons to zoom out or in on the graph. Click the "Lighten" and "Darken" buttons (or enter a value in the box and press Enter) to change how dark a particular score point appears. Click one of the Window buttons to change the length over which the score displayed is averaged. A window of length N will produce dots that represent the score over the N residues up to the dot display point. A higher N value will tend to increase the contrast in regions of good local alignment but may suppress short local alignments and make chance local matches appear more prominent.

Because the global scores for most alignments will be much higher than the local scores (because they accumulate over the length of the alignment) you will want to view the global scores with a much lower gray scale. The minimum setting is 0.01 and the maximum is 100.

See section at end to understand what the dot matrix display can tell you.

Highlighting alignment path: use the menu "Display : Current Path" to show the optimal alignment path through this dot matrix, which is the one shown in the standard and text alignment panels. The path shows up as a translucent green line.

Selecting sequence: Ctrl-click on the dot matrix, drag, and release. A pink box delineates the dot matrix space covered and the specific residues in the two sequences will be displayed.

Aligning selected sequence: Use the menu "Align : Align Selected Regions" to produce a new alignment of the selected regions. You can reset the new alignment's pair align settings using the menu "Settings : Pair Align Settings".

Showing cursor position: Alt-click on the dot matrix to show guide lines and the exact position of the click in the two sequences.


File Menu:

Save as JPEG: Save this dot matrix image as a JPEG format image file.

Print: Print this dot matrix image. This is simple so far - no scaling or labeling. If your display is large it will run off the page. If so, reduce the scale on screen before printing.


Align Menu:

Align Selected Region: Generate a new alignment from the region that is selected on the dot matrix (Ctrl-click and drag to select a region).


View Menu:

Current Path: Display the best alignment path that was determined for this alignment.

Reset Panel Scale: Sequences can be selected for alignment only with an unmagnified dot matrix image. Use this to quickly return to this scale.

Clear Selections: Clear extraneous labels and selections.

Local Scores: Display dots that represent the local alignment quality.

Global Scores: Display dots that represent the best alignment score up to each dot position.


Settings Menu:

Pair Alignment Settings: Show the Pair Alignment Settings frame to change settings for further alignments initiated from this frame.


Help Menu:

Dot Matrix Help: This help.


Understanding the dot matrix: The dot matrix display can be a powerful way of viewing a sequence alignment. It is especially revealing under these conditions: 1) there are repeating domains in the proteins, 2) well-aligned regions are scattered sparsely in the proteins, 3) well-aligned regions appear in unexpected places (e.g. out of order), and 4) there a prominent near-optimal alignments. The reason it can be so useful is that the dot matrix shows you the score results for an entire dynamic programming alignment, rather than showing only the single best alignment. Conceptually, one sequence defines the X-axis, N-terminus at the left, and the other sequence defines the Y-axis, N-terminus at the top (see Figure 1 below). The left-most vertical band (by default, one pixel wide) represents the first amino acid from sequence A aligned with every amino acid from sequence B (running from top to bottom). The next vertical band represents the same alignment with the second amino acid from sequence A, etc. until the final vertical band represents the last amino acid from sequence A aligned with every amino acid from sequence B. At each point on the graph, the darkness of the dot represents the score for the corresponding two residues when aligned, or a local summation of scores for several aligned residues leading up to that residue (see below).

Figure 1

Two sequences that are similar across their entire length, as in Figure 1 above, will have a prominent diagonal line running from the top left (where their N-termini align) to the bottom right (where their C-termini align). This reflects a relative density of high-scoring aligned residue pairs along their length. The line will be patchy or jogged if there are regions of low sequence similarity or gaps (indels). An example of a jog corresponding to a sizable gap in sequence B is found in Figure 1 in the lower right region. That section is shown magnified in Figure 2 (below), with red highlights under the ungapped diagonal match regions and a green dashed line showing where the jog occurs, corresponding to a gap in sequence B. The patchy jogging diagonal line will usually correspond approximately to the optimal alignment between the two proteins, which can be viewed in the standard alignment frame. In the dot matrix view, use the menu "Display : Current Path" to highlight the optimal alignment path through the dot matrix. In most cases, you will see both horizontal and vertical jogs in the highlighted diagonal that represent regions of gaps in one sequence or the other (think of these as regions where one sequence continues along its axis while the gapped sequence pauses on its axis). It is instructive to select a few such regions, including a little surrounding well-matched sequence, and realign them with "Allow Local Alignment" turned OFF in Pair Alignment Settings. In the standard alignment window you should see regions of good alignment with a prominent region of indels between.

Figure 2

The dot matrix is especially valuable in providing a visual overview of all possible alignments. Internal repeats will show up as diagonals off the main diagonal (this is a good reason to align a sequence to itself and view the dot matrix). Even subtle features of local alignment may be visible. An example is shown below, in which the R/KXX repeats of the voltage sensor from two related K-channels can be seen as off-diagonal echoes spaced 3 residues off the main diagonal (this is a very magnified view - each residue pair with a favorable score is visible as a dark spot so that you can count positions readily).

Figure 3


James H. Thomas, Department of Genome Sciences, University of Washington
8/1/2002