‹‹ Back to SVS Home

Using LD Plots

[next] [prev] [prev-tail] [tail] [up]

11.9 Using LD Plots

Linkage Disequilibrium (LD) is the non-random association of alleles in a population. LD is useful during analysis to identify linkage relationships of interesting markers. LD can be calculated for any spreadsheet with more than one Genotypic type data column. Additionally, the LD visualization provides a means to interactively define haplotype blocks for haplotype association testing (see Haplotype Association Tests).

LD Plot Modes

LD can be visualized with or without marker map data. When marker map data is available Plot Viewer will present the LD data in position scaled form and enable the Genome Browser.

Otherwise, the data will be presented in a uniform scale, and without the Genome Browser.

[Picture]

Figure 93: Comparison of a position scale LD visualization (left), and a uniform scale (right).

The markers are the individual pentagons along the spine of the LD graph. The width of a marker is determined using the midpoint between a marker and its adjacent markers, thus the exact marker position is always within the marker representation, but may not be centered. Edge markers, however, use symmetry to define the open side and so in those cases the marker is centered within the representation.

Graph Controls Specific to LD Plots

On the Graph tab, the controls specific to LD are:

  • Vertical Flip text
  • Overlay
  • Marker color
  • LD color

The Vertical Flip check box allows for the specification of the location of the spine of the visualization to be either along the top or the bottom of the graph. This control is checked by default, corresponding to the spine at the top of the graph.

The Overlay check box allows for a grid to be drawn over the marker and LD elements. The Overlay is only drawn for markers that are wide enough that drawing the overlay will not obstruct the data. In some position scaled visualizations, there may be some markers that are too wide to draw the overlay, and some that are not. By zooming into a dense region, the marker representations can become large enough to draw the overlay. Unchecking the control will disable this feature entirely.

The Marker color button provides a color chooser for the marker representations.

On the Color tab, color buttons define the colors for the default values of 0.0, 0.5, and 1.0 for Min, Mid, and Max, respectively. These values can be changed by double clicking on the values. Additional splits can be added by clicking the Add button. Splits can be removed by selecting a split and clicking Remove. The minimum number of splits allowed is two. There is no maximum number of splits. Colors for other values are computed by interpolating between the color values selected.

On the Method Options tab, the options for setting LD parameters are available. These options are:

  • Method: The method consists of two parts, the method and the statistic, the first menu is the method (EM or CHM) and the second is the statistic (R2 or D′)
  • Options for EM Method: (Only available when EM method is selected.)
    • Use samples with missings: Indicates whether samples with missing values should be used for the computation.
    • Max Iterations: The maximum number of iterations used for EM computation.
    • Conv Tolerance: Convergence Tolerance for EM method computation.
Haplotype Block Sets and LD Graphs

The Haplotype Block Set item is a fixed item of an LD graph, i.e. it cannot be removed, nor can another one be added to the graph. The Block Definitions section of the control panel for blocks presents the following four buttons:

  • Load...
  • Save
  • Clear All
  • Compute...

The Load button activates a spreadsheet selection dialog. A haplotype block definition spreadsheet can then be selected. If the selected spreadsheet contains multiple columns that could be block definitions, an additional dialog will appear for selecting the correct column.

The Save button creates a new spreadsheet from the marker definitions. The button is only enabled if blocks are created of modified manually. If manual changes are made and the save button is not pressed before closing Plot Viewer, a prompt will ask whether the haplotype block changes should be saved or not.

The Clear All button removes all blocks from the graph.

The Compute button activates the Haplotype Block Detection dialog using either all the data available in the plot, or just the data currently visible in the plot. When the dialog’s Run button is pressed, a output spreadsheet will be created as a child of the genetic spreadsheet before it is loaded into the viewer.

Tables for Selected Block

A block can be selected by clicking anywhere within a block on the graph. When a block is selected, the Tables for Block # section will be enabled. The Haplotype Tables button will activate the Haplotype Tables dialog. See Haplotype Tables for more information. Pressing Run will generate spreadsheets in the project navigator as a children of the current Plot Viewer node.

The Subset Markers button performs a column subset of the genetic spreadsheet, where the columns correspond to the markers from the selected block as well as any non-genetic columns from the original spreadsheet. The output spreadsheet is a child of the genetic spreadsheet.

Frequency Threshold for Console Output

When a block is selected, the application will begin computing haplotype statistics. When the data is ready, it will be presented in the Data Console. To limit the volume of content presented in the console, a haplotype frequency must exceed the Frequency Threshold value. Acceptable values are between 0.00001 and 0.999.

Interactive Block Definition

Haplotype blocks can be created, modified, or deleted via a menu which is accessible by right-clicking on a marker. Create Block ... will be available when clicking on a marker that is not already a member of another block. This action will create a new block with one marker. It will be labeled with the next available number in the block set. A block can be modified by adding or removing markers. Markers adjacent to the edge of the block can be included or excluded by clicking and dragging the vertical line segment of the haplotype block. Markers can be dropped from a block by selecting Remove ... from the menu. Dropped markers from the middle of a block can be added back in by selecting Add ....

The Full Domain View in LD Plots

The Full Domain view for LD graph shows an average value of the LD for adjacent markers. It is not intended as an analytic resource. This view is useful to provide a sense of context while inspecting the graph more closely.

The Data Console in LD Plots

The Data Console provides a detailed text output including hyperlinks when left-clicking on graph elements. There are 3 elements comprising an LD graph:

  • Markers
  • LD Values
  • Blocks Definitions

Left-clicking a marker produces the marker name, it’s position (if it is position scaled), it’s minor allele frequency, and several hyperlinks to various helpful websites.

Clicking an LD value produces the marker information for both markers, distance calculations, computed LD values, and (if available) upper and lower confidence bounds of the computation.

Clicking on a haplotype block unavoidably includes clicking on a marker or an LD value. Clicking will first produce the data for the marker or LD value, followed by block information including its genomic location (if available) the block width, and several useful website links. Lastly, haplotypes with frequencies above the frequency threshold set in haplotype block control panel are listed in decreasing frequency order.

Adding Additional Graphs

In the case that data is marker mapped (and therefore position scaled) the Add Graphs control item will be visible when clicking on the User Graphs top-level node item. Selecting it will make the Add Graph control panel accessible. See Using the User Graphs Tree Window for more information on how to add graphs.

[next] [prev] [prev-tail] [front] [up]