ComparativeMarkerSelectionViewer (v7)

Views the results from ComparativeMarkerSelection

Author: Joshua Gould

Contact:

gp-help@broadinstitute.org

Algorithm Version:

The ComparativeMarkerSelectionViewer provides tools for reviewing and working with ComparativeMarkerSelection results. These tools include plots, a data table, expression profiles and heatmaps, filters, annotations, and creating dataset and feature list files. (A feature list is a text file with one feature per line.)

ComparativeMarkerSelectionViewer is an interactive tool. For non-interactive access to annotations, heatmaps, and creation of datasets and features lists, see the HeatMapImage, and ExtractComparativeMarkerSelection modules.

Plots

Plots are displayed in the upper portion of the viewer. Use the Window menu to select the plot to display:

  • Upregulated Features (initially displayed) plots the descending test statistic values versus features sorted by rank. This plot is useful for visualizing the number of features that have a positive and negative test statistic value in each class.
  • Comparison plots (such as FDR(BH) vs Q Value) provide a pair-wise comparison of different significance measures. These plots are useful to assess the relative stringency of the selected hypothesis rejection criteria. To display the comparison plots, click Window>FDR(BH) vs Q Value. To compare different significance measures: use the drop-down lists below the plot to select the measures to display on the X and Y axes and click Change Axes. The viewer updates the plot and the Window menu to reflect the new comparison.
  • λ vs. π0 plots the estimated π0 versus the tuning parameter λ. Use this plot to see how well the spline fits the data; that is, the quality of the π0 estimate (1).
  • Histograms show the null distributions for each measure of significance.
  • General Information lists the parameters used to generate the ComparativeMarkerSelection results.

Use the following menu items to work with any plot:

View>Zoom In
View>Zoom Out
Zoom in and out on the plot. (Alternatively, use your mouse to click-and-drag over the area to zoom in on.)
View>Reset Return to the default zoom level.
View>Display Options Change the title, labels, and other visual attributes.
File>Save Image Save the plot to an image file.
File>Print Print the plot (formats the plot to fit the page).

Data Table

The table in the lower portion of the viewer lists the ComparativeMarkerSelection results. For a description of each column, see the ComparativeMarkerSelection documentation. By default, the features in the table are ordered by score. To reorder the features, click the title of the key column; for example, to order features by rank, click Rank.  

To choose which columns are shown in the table, click the Column Chooser icon in the top-right corner of the table, just above the vertical scrollbar:

Using File>Export Current Table as CSV, you can export the table for use in external programs such as Excel.  This will result in a comma-spearated values (CSV) file with the contents of the table, leaving out any items excluded due to filters or columns hidden through the Column Chooser.

Expression Profiles and Heatmaps

The expression profile for a feature plots expression value per sample. A heatmap shows the same information, but color codes the expression values from red for the highest expression values through blue for lowest expression values. 

To display an expression profile or heatmap:

  1. Select one or more features from the data table.
  2. Click View>Profile or View>Heatmap. Alternatively, right-click the data table and select Profile or Heatmap from the context menu.

If no features are selected, all features are displayed. Ordering all features by score and displaying them in a heatmap can be very informative. Displaying all features in a profile can be time consuming and is less useful.

Note: If the Profile and Heatmap commands are unavailable, your dataset is not open. To open your dataset, specify the dataset filename parameter when you launch the ComparativeMarkerSelectionViewer or, from the viewer, select File>Open Dataset.

Filters

Use filters to display only features that you are interested in:

  1. Select Edit>Filter Features>Custom Filter. The Filter Features window appears.
  2. In the Filter Features window, set your filtering criteria and click OK. The viewer shows only those features that meet your criteria.
  3. To redisplay all features, select Edit>Filter Features>Show All.

Creating Datasets/Feature Lists

You can use the viewer to create a new dataset or feature list from the ComparativeMarkerSelection results:

  1. Select File>Save Derived Dataset or File>Save Feature List. A window appears.
  2. Choose the features to include in the dataset/feature list:
    • Use current features – the features currently displayed in the viewer (use filters to display the features that you want in the new dataset/feature list).
    • Use all features – all features (ignore any filters).
    • Use selected features – the features currently selected in the data table.
  3. Choose a location and name for the new dataset/feature list.
  4. Click Create to save the new data set/feature.

Note: If the Save Derived Dataset command is unavailable, your dataset is not open. To open your dataset, specify the dataset filename parameter when you launch the ComparativeMarkerSelectionViewer or, from the viewer, select File>Open Dataset.

Annotations

The ComparativeMarkerSelectionViewer provides two annotation methods:

  • Feature annotations use color to annotate features in the data table
  • GeneCruiser annotations retrieve information about Affymetrix probe ids and add it to the data table  
    Support for GeneCruiser ended in March 2016
     

To use feature annotations:

  1. Create a feature list file (one way of doing this is described above).
  2. Select File>Open Feature List(s) to open your feature list file.  In the Feature column of the data table, a color bar appears next to each feature in the feature list.
  3. Select Edit>Feature Annotations to edit the color or close the feature list:
  4. In the Feature Annotations window, select your feature list from the drop-down list. The color assigned to that feature list appears in the box to the right.
    • To change the color, click the box and select a new color.
    • To close the feature list and remove the color bars from the table, click Delete.

To use GeneCruiser annotations:

  1. Select GeneCruiser>Gene Information.
  2. Select the features that you want to retrieve annotations for in the table.
  3. Choose which fields to retrieve from GeneCruiser in the GeneCruiser dialog.
  4. The annotations appear in additional columns in the table.

References

Gould et al. Comparative Gene Marker Selection suite. Bioinformatics Advance Access published on May 18, 2006, DOI 10.1093/bioinformatics/btl196.

Parameters

Name Description
comparative marker selection filename * The output from ComparativeMarkerSelection
dataset filename  The dataset file used as input to ComparativeMarkerSelection

* - required

Input Files

  1. comparative marker selection filename
    The output file from ComparativeMarkerSelection in ODF format.
  2. dataset filename
    The dataset file used as input to ComparativeMarkerSelection in GCT or RES format.

Platform Dependencies

Task Type:
Visualizer

CPU Type:
any

Operating System:
any

Language:
Java

Version Comments

Version Release Date Description
7 2015-03-17 Fix for SecurityException when saving datasets
6 2014-07-28 Added filter by multiple criteria, hide table columns, export table. Fixed for use with Java 7. Many other improvements and bug fixes.
5 2013-09-20 Added filter by rank, filter by multiple criteria, hide table columns, export table. Fixed for use with Java 7. Many other improvements and bug fixes.
4 2006-07-29 Updated GeneCruiser library
3 2006-03-09 Updated to work with latest ComparativeMarkerSelection module
2 2005-06-13 Added maxT p-value and removed rank p-value